{"$schema": "https://c3voc.de/schedule/schema.json", "generator": {"name": "pretalx", "version": "2026.1.0.dev0"}, "schedule": {"url": "https://pretalx.com/juliacon-local-paris-2025/schedule/", "version": "0.5", "base_url": "https://pretalx.com", "conference": {"acronym": "juliacon-local-paris-2025", "title": "JuliaCon Local Paris 2025", "start": "2025-10-02", "end": "2025-10-03", "daysCount": 2, "timeslot_duration": "00:05", "time_zone_name": "Europe/Paris", "colors": {"primary": "#4063d8"}, "rooms": [{"name": "Coffee room", "slug": "4000-coffee-room", "guid": "5371b68f-4cd8-50a4-8987-624d923e6ad2", "description": "Above both amphitheaters", "capacity": null}, {"name": "Jean-Baptiste Say Amphitheater", "slug": "3998-jean-baptiste-say-amphitheater", "guid": "333d52de-f5a3-5950-a627-aaa15159b984", "description": null, "capacity": 175}, {"name": "Robert Faure Amphitheater", "slug": "3999-robert-faure-amphitheater", "guid": "f9faa2f5-72cd-5411-b9ae-b9ec8ae094fd", "description": null, "capacity": 175}], "tracks": [{"name": "Language & teaching", "slug": "6202-language-teaching", "color": "#940721"}, {"name": "Neuroscience", "slug": "6207-neuroscience", "color": "#f4f116"}, {"name": "SciML", "slug": "6204-sciml", "color": "#019108"}, {"name": "Quantum", "slug": "6208-quantum", "color": "#f4f116"}, {"name": "Numerical analysis", "slug": "6203-numerical-analysis", "color": "#0c5289"}, {"name": "Applications", "slug": "6209-applications", "color": "#e09c14"}, {"name": "Machine learning", "slug": "6205-machine-learning", "color": "#d00000"}, {"name": "Randomness", "slug": "6200-randomness", "color": "#0b728e"}, {"name": "GPU & performance", "slug": "6201-gpu-performance", "color": "#b717ed"}, {"name": "Optimization", "slug": "6206-optimization", "color": "#00c3f6"}, {"name": "Posters", "slug": "6216-posters", "color": "#a2a2a2"}, {"name": "Error, derivatives, stability", "slug": "6210-error-derivatives-stability", "color": "#e20bb0"}, {"name": "General", "slug": "6211-general", "color": "#000000"}], "days": [{"index": 1, "date": "2025-10-02", "day_start": "2025-10-02T04:00:00+02:00", "day_end": "2025-10-03T03:59:00+02:00", "rooms": {"Jean-Baptiste Say Amphitheater": [{"guid": "7948f114-dcbb-5215-a9f8-18aa1889136e", "code": "UWYT8L", "id": 79886, "logo": null, "date": "2025-10-02T08:50:00+02:00", "start": "08:50", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79886-opening", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/UWYT8L/", "title": "Opening", "subtitle": "", "track": "General", "type": "Lightning talk", "language": "en", "abstract": "-", "description": "-", "recording_license": "", "do_not_record": false, "persons": [{"code": "PJVBLC", "name": "Guillaume Dalle", "avatar": "https://pretalx.com/media/avatars/PJVBLC_t336z2c.webp", "biography": null, "public_name": "Guillaume Dalle", "guid": "2515531a-5214-559d-968e-e8e0d2adc883", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PJVBLC/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UWYT8L/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UWYT8L/", "attachments": []}, {"guid": "01d63a2d-22a9-533e-9e75-01d87db88374", "code": "8YKBWV", "id": 79775, "logo": null, "date": "2025-10-02T09:00:00+02:00", "start": "09:00", "duration": "01:00", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79775-randomization-for-solving-high-dimensional-problems-algorithms-and-software", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/8YKBWV/", "title": "Randomization for solving high-dimensional problems: algorithms and software", "subtitle": "", "track": "Randomness", "type": "Plenary", "language": "en", "abstract": "Randomization is a powerful dimensionality reduction technique that allows to solve large scale problems while leveraging optimized kernels and enabling the usage of mixed precision. In this talk we will review recent progress in using randomization for solving linear systems of equations or eigenvalue problems. We first discuss sketching techniques that allow to embed large-dimensional subspaces while preserving geometrical properties and their parallel implementations. We then present randomized versions of processes for orthogonalizing a set of vectors and their usage in the Arnoldi iteration. We discuss associated Krylov subspace methods for solving large-scale linear systems of equations and eigenvalue problems. The new methods retain the numerical stability of classic Krylov methods while reducing communication and being more efficient on modern massively parallel computers. We finally discuss their implementation in a Julia library.", "description": "Laura Grigori is a Professor at EPFL and former Director of Research at Inria, where she led the Alpines team. Her research focuses on numerical linear algebra, communication-avoiding algorithms, and scalable solvers for scientific computing. She has made influential contributions to high-performance computing with applications in physics and engineering.", "recording_license": "", "do_not_record": false, "persons": [{"code": "HQTQPU", "name": "Laura Grigori", "avatar": null, "biography": "Laura Grigori is a Professor at EPFL and former Director of Research at Inria, where she led the Alpines team. Her research focuses on numerical linear algebra, communication-avoiding algorithms, and scalable solvers for scientific computing. She has made influential contributions to high-performance computing with applications in physics and engineering.", "public_name": "Laura Grigori", "guid": "985a8f11-f9a3-54a3-a7fb-fc75519e14af", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HQTQPU/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/8YKBWV/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/8YKBWV/", "attachments": []}, {"guid": "84b551e4-e750-566f-aa3b-333f95019079", "code": "Q7G3C3", "id": 70534, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/Q7G3C3/rmt_talk_ujXG6y8.png", "date": "2025-10-02T10:00:00+02:00", "start": "10:00", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70534-interactive-random-matrix-theory-in-julia", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/Q7G3C3/", "title": "Interactive Random Matrix Theory in Julia", "subtitle": "", "track": "Randomness", "type": "Talk", "language": "en", "abstract": "What do growth from a corner, random domino tilings, nonintersecting paths and Young tableaux generated from random words have in common? That they can all be simulated in Julia of course! This talk will present various Julia packages for the statistical analysis of determinantal point processes just like these with a focus on interactive exploration through tools such as Pluto.jl and WGLMakie.jl.", "description": "Determinantal point processes (DPPs) have attracted attention in machine learning recently as probabilistic models that naturally enforce some diversity in a set. They also occur in many areas of mathematics and mathematical physics, connecting processes such as corner growth, the statistics of random domino as well as lozenge tilings or randomly generated Young tableaux.\r\n\r\nWhat are DPPs? At their core, they describe a probability distribution on subsets using matrix determinants. This gives rise to various interesting phenomena and associates all the processes described above with a matrix ensemble. Rather than focusing on the theoretical aspects of these processes however, this talk will dive right into the simulation and experimental analysis of DPPs. Being able to interactively engage with these processes through computation not only is a powerful teaching tool but also opens the door to further research.\r\n\r\nScaling such simulations is challenging due to the nontrivial amount of computation required to generate large enough statistics. This is where a JIT-compiled language like Julia shines, allowing one to take full advantage of the hardware and providing easy access to thread-level parallelism as well as GPU computing, while still providing an interactive environment for exploratory analysis. These advantages transfer to numerous applications outside of random matrix theory as well, making DPPs a good case study for interactive computation with demanding performance requirements.\r\n\r\nBased on the example of AztecDiamonds.jl, I demonstrate the ability to scale simulations from running serially on a CPU to employing KernelAbstractions.jl to run the domino shuffling algorithm in parallel on GPUs, thus making generating much larger diamond tilings than ever before feasible. For inspecting results, the package integrates nicely with the Images.jl ecosystem as well as Makie.jl, which can be used to create interactive Pluto.jl experiences perfectly taylored to a classroom setting.\r\n\r\nI will give an overview over existing packages that are useful for simulating DPPs as well as ones I created based on the requirements of dealing with random Young tableaux, discrete polynomial ensembles and random lozenge tilings and present educational resources created using those libraries. By making these tools openly available, I hope to enable and inspire further research into the subject. Lastly, I will also give an outlook into my own ongoing research into rhombus tilings of polygons and how I am using Julia to gain a better understanding of the subject.", "recording_license": "", "do_not_record": false, "persons": [{"code": "LWE98T", "name": "Simeon Schaub", "avatar": "https://pretalx.com/media/avatars/LWE98T_R3EkBeu.webp", "biography": "Master's student at KIT (Germany)\r\n\r\nInterested in computational mathematics, programming language design, automatic differentiation and compilers.\r\n\r\nGitHub: https://github.com/simeonschaub", "public_name": "Simeon Schaub", "guid": "227c2d9c-9197-5f2f-9862-e26dd61967fc", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/LWE98T/"}], "links": [{"title": "The AztecDiamonds.jl package", "url": "https://github.com/JuliaLabs/AztecDiamonds.jl", "type": "related"}], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Q7G3C3/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Q7G3C3/", "attachments": []}, {"guid": "d18bf753-1546-5947-856c-5737b7f3d826", "code": "UDQLTD", "id": 72251, "logo": null, "date": "2025-10-02T10:30:00+02:00", "start": "10:30", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-72251-stochastic-rounding-when-0-1-0-2-0-3-does-equal-zero-at-least-on-average", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/UDQLTD/", "title": "Stochastic Rounding: When 0.1 + 0.2 - 0.3 does equal zero (at least on average)", "subtitle": "", "track": "Randomness", "type": "Lightning talk", "language": "en", "abstract": "In Float64, 0.1 + 0.2 \u2013 0.3 is not zero due to round-off errors. Try it. Stochastic rounding provides a trick: floating point rounding is done probabilistically: sometimes up, sometimes down. This allows for numerical simulations on hardware made for machine learning, where Float16 precision is common. We can perform meaningful simulations even with limited bits.\r\nAnd yes, all of this can be done in Julia.", "description": "In standard floating point (FP) arithmetic, the simple expression 0.2 + 0.1 - 0.3 never evaluates to zero. Try it now in your REPL. This happens due to numerical round-off errors, as specified by IEEE 754. But what if the expression is zero on average? This is where stochastic rounding (SR) comes into play.\r\n\r\nIn a nutshell, SR decides after each FP operation whether to round up or down to fit the exact result into the FP type. The probability depends on the distance to the next available larger or smaller FP. The topic reappeared recently in research in the context of machine learning. New specialized hardware can perform huge parallel computations in low precision, down to Float16.\r\n\r\nSR compensates for numerical errors in low precision, making it applicable to numerical simulations, as shown in [1,2]. StochasticRounding.jl [3] provides a Julia implementation of stochastic FP types, where the effects of stochastic rounding can be investigated. However, since this is done in software, it is slow. Nevertheless, a first processor unit  that performs SR in silicon is available [4]. And yes, we can run Julia code on such machines [5]. There is industrial interest, and other vendors may follow to implement SR in their hardware.\r\n\r\nThis talk will provide an overview of the mathematical properties of SR, the implementation in Julia, as well as showcase the potential of running large-scale simulations at low precision.\r\n\r\n\r\n[1] Kl\u00f6wer, M. et al. Periodic orbits in chaotic systems simulated at low precision. Sci Rep 13, 11410 (2023). https://doi.org/10.1038/s41598-023-37004-4\r\n\r\n[2] Paxton, E. et al. Climate Modeling in Low Precision: Effects of Both Deterministic and Stochastic Rounding. Journal of Climate, 2022, http://dx.doi.org/10.1175/JCLI-D-21-0343.1\r\n\r\n[3] https://github.com/milankl/StochasticRounding.jl\r\n\r\n[4] https://www.graphcore.ai/products/ipu\r\n\r\n[5] https://giordano.github.io/talks/2024-06-20-julia-ipu-ornl", "recording_license": "", "do_not_record": false, "persons": [{"code": "HVFDHH", "name": "Patrick Jaap", "avatar": "https://pretalx.com/media/avatars/HVFDHH_4dKMvnl.webp", "biography": "I'm a postdoc researcher at the Weierstrass Institute in Berlin in the scientific computing group.\r\nMy day-to-day work is all about running finite element simulations for solid mechanics and quantum devices.\r\nI came into touch with Julia two years ago and I love it since.\r\nAt WIAS, we have our own Julia PDE solver ecosystem WIAS-PDELib, where I am a core maintainer.\r\nI'm interested in low-level solver routines, software architecture, coding quality, and experiments with new data types and programming features.", "public_name": "Patrick Jaap", "guid": "c7449cb3-36a0-5883-a1c6-4810c0b1ba47", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HVFDHH/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UDQLTD/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UDQLTD/", "attachments": []}, {"guid": "e350b952-f2a6-5ff9-baff-cbd37b8a4741", "code": "YKP9ZD", "id": 68020, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/YKP9ZD/logo_XIVUbWe.png", "date": "2025-10-02T10:40:00+02:00", "start": "10:40", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-68020-stochasticweathergenerators-jl-a-package-to-generate-la-pluie-et-le-beau-temps", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/YKP9ZD/", "title": "StochasticWeatherGenerators.jl \u2014 a package to generate La pluie et le beau temps!", "subtitle": "", "track": "Randomness", "type": "Lightning talk", "language": "en", "abstract": "Climate change challenges demand assessing the resilience of key sectors, such as energy and agriculture, to future weather conditions.\r\nStochastic Weather Generators (SWGs) are essential tools for this purpose, enabling efficient sampling of climate statistics\u2014especially extreme events.\r\n\r\nCombined with impact models like crop models, SWGs can be used to stress-test and measure system sensitivity to climate variability, particularly under climate change. For example, estimating the frequency of long dry spells is crucial.\r\n\r\nThe [StochasticWeatherGenerators.jl](https://github.com/dmetivie/StochasticWeatherGenerators.jl) package implements several SWG models, including those based on Hidden Markov Models and GeoStatistics functions.\r\nIts goal is to provide a range of models and tools to facilitate comparison and experimentation.\r\n\r\nThanks to Julia\u2019s powerful ecosystem and capabilities, defining, fitting, and using stochastic weather generators has never been easier!", "description": "Many Stochastic Weather Generators (SWGs) models have been proposed over the years, but their implementations have largely remained private. In recent years, a few GitHub repositories have appeared\u2014mostly R code (often with C++ parts)\u2014but typically without documentation or with very complex instructions.\r\nThis has made the SWG ecosystem sparse and misaligned with modern reproducibility practices, often preventing comparisons between models.\r\n\r\nIn this talk, I will present my attempt at a unified Stochastic Weather Generator package.\r\nSWGs generally rely on advanced statistical frameworks (HMMs, Bayesian models, geostatistics, copulas, etc.). While most researchers in the community are bound to the R ecosystem (with the usual two-language problem), this package is built using the Julia ecosystem and benefits from what has already been implemented elsewhere.\r\n\r\nAlongside the package, I will showcase a few applications in climate and agronomy.\r\n\r\nNote: this is still very much a Work in Progress and I will be happy to take your suggestion, in particular regarding the organization of the package!", "recording_license": "", "do_not_record": false, "persons": [{"code": "GLXYLX", "name": "David M\u00e9tivier", "avatar": null, "biography": "I am permanent researcher at INRAE Montpellier since 2023. I work in Applied Mathematics and Physics. My latest research interests focus on statistical models e.g. Hidden Markov Models, Deep Learning to tackle environmental, and climate (change) problems. I also work on practical Robust statistics. Previously, I did a PhD and half a postdoc in theoretical physics on mean field dynamics of particles systems, looking at as bifurcations, synchronization, instabilities, partial differential equations, etc. I am a Julia enthusiast, and I am working on a few packages (see my GitHub profile.).", "public_name": "David M\u00e9tivier", "guid": "fb534dc0-905c-50af-b77e-303bbcf1cd8d", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/GLXYLX/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/YKP9ZD/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/YKP9ZD/", "attachments": []}, {"guid": "2e2e70b5-e276-56f2-aa32-7b899c211816", "code": "KWGKGS", "id": 70660, "logo": null, "date": "2025-10-02T10:50:00+02:00", "start": "10:50", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70660-streamsampling-jl-efficient-streaming-sampling-methods-in-julia", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/KWGKGS/", "title": "StreamSampling.jl: Efficient Streaming Sampling Methods in Julia", "subtitle": "", "track": "Randomness", "type": "Lightning talk", "language": "en", "abstract": "Sampling is at the base of many statistical methods and procedures. Apart from the traditional sampling from finite populations available in StatsBase.jl, there are many more methods and techniques which could be of interest in specific simulation scenarios where performance and memory consumption are critical. In such cases, packages implementing sampling methods for streaming data could be particularly advantageous for scalable and efficient simulations.", "description": "For applications that handle high\u2010velocity data as real\u2011time monitoring of network traffic or sensor feeds in IoT deployments, the classical paradigm of drawing all observations into memory before sampling can become infeasible. Streaming algorithms overcome this bottleneck by maintaining only a compact summary of the data seen so far and updating the sample incrementally as each new element arrives.\r\n\r\nWe will present a package which expands the sampling tools in Julia in this direction: StreamSampling.jl. The package focuses on techniques for sampling from large or unbounded streams of data which could be too large to store in memory, providing methods such as reservoir sampling and other streaming algorithms that ensure efficient and representative sampling in this setting. \r\n\r\nInterestingly, by eliminating the need for an initial memory-allocating pass to collect the elements, some of these methods have also shown performance benefits when sampling from generic iterators, which would otherwise need to be stored in memory if a traditional approach had been used.", "recording_license": "", "do_not_record": false, "persons": [{"code": "YVBFVK", "name": "Adriano Meligrana", "avatar": null, "biography": "I\u2019m a Junior Researcher at the CENTAI Institute, specializing in statistical and simulation methodologies.", "public_name": "Adriano Meligrana", "guid": "eb47bcb4-94e9-5d93-b7f6-78fb81247d7b", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/YVBFVK/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KWGKGS/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KWGKGS/", "attachments": []}, {"guid": "72350550-ca35-51d2-8694-7194a7070917", "code": "7KZ9LX", "id": 79778, "logo": null, "date": "2025-10-02T11:30:00+02:00", "start": "11:30", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79778-beating-the-market", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/7KZ9LX/", "title": "Beating the market", "subtitle": "", "track": "General", "type": "Sponsor talk", "language": "en", "abstract": "At Temple Capital we empower researchers to rapidly prototype over a massive design space and deploy what works. Our backtesting and walkforward analytics are fast and reliable in part because we deploy the very same code and datastructures that researchers prototype with to production in a live trading environment. This talk will share how we leverage Julia to build re-usable datastructures and expressive, flexible, and performant paradigms to find alpha and beat the market.", "description": "Sponsor talk by Temple Capital (https://www.temple.capital/).", "recording_license": "", "do_not_record": false, "persons": [{"code": "UYMQJS", "name": "Lilith Hafner", "avatar": "https://pretalx.com/media/avatars/UYMQJS_nFo0m83.webp", "biography": null, "public_name": "Lilith Hafner", "guid": "68e94df8-a99a-5d38-84d8-145e7ef0b643", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/UYMQJS/"}, {"code": "YS9C7X", "name": "Vasily Kokorev", "avatar": null, "biography": null, "public_name": "Vasily Kokorev", "guid": "f913a759-315d-56d4-a576-36173f793205", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/YS9C7X/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7KZ9LX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7KZ9LX/", "attachments": []}, {"guid": "80eb72d6-d8f7-560e-a44a-6a77590b4661", "code": "F3GZJM", "id": 81646, "logo": null, "date": "2025-10-02T11:40:00+02:00", "start": "11:40", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-81646-bringing-flight-recorder-tech-to-julia-and-building-its-future-together", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/F3GZJM/", "title": "Bringing Flight Recorder Tech to Julia \u2014 And Building Its Future Together", "subtitle": "", "track": "General", "type": "Sponsor talk", "language": "en", "abstract": "Seven years ago, I hit my first \u201cbug from hell\u201d, the kind of issue that would be trivial to solve if we had something like an airplane\u2019s flight data recorder, but for software. A system that records every change in execution so you can replay and analyze issues down to the exact line of code.\r\nThat idea grew into what we now call CodeGlass: technology that provides time-travel analysis and deep runtime insights, already helping enterprises tackle elusive bugs, performance bottlenecks, and memory issues across multiple programming languages.\r\n\r\nLast year, members of the Julia community asked us: <i>can we bring this to Julia?</i>\r\n\r\nIn this talk, I\u2019ll explain how the technology works, where it can help Julia developers today, especially in high-performance computing, scientific research, and reliability, and what it could enable next for the Julia community, like trace-based debugging. Most importantly, we want to connect with researchers, engineers, developers, and organizations who can benefit from this now, and collaborate with us to shape the future of CodeGlass in the Julia ecosystem.", "description": "Sponsor talk by CodeGlass (https://codeglass.io/).", "recording_license": "", "do_not_record": false, "persons": [{"code": "PSVYRC", "name": "Tyrone Krieger", "avatar": null, "biography": "I've been writing code since I was 11. Nearly two decades later, I'm still baffled by the fact that most developers spend only 32% of their time actually coding.\r\nMy professors used to say this was just the way things were. But instead of accepting it, I decided to push back. One step at a time.\r\nWhy? Because we can.\r\nAs developers, we build the tools that move entire industries forward. So why not turn that same energy inward and improve our own?\r\nStill think it's impossible? Have a look: https://CodeGlass.io\r\n\r\nWhat I Love:\r\n\u2022\tDiving deep into complex codebases\r\n\u2022\tSharing developer knowledge\r\n\u2022\tBuilding powerful tools (like CodeGlass)\r\n\u2022\tExploring superconductors and the Meissner effect (hoverboards when?)\r\n\u2022\tI Like Trains\r\n\u2022\tLizard Doggo \r\n\r\nLet\u2019s grab a digital drink: Tyrone@CodeGlass.io or Discord #aBrokenDonut", "public_name": "Tyrone Krieger", "guid": "212edd21-a1f7-53d0-b700-8778df37462b", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PSVYRC/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/F3GZJM/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/F3GZJM/", "attachments": []}, {"guid": "19bfa749-8608-54dd-97e1-dc07f69d9b02", "code": "ZLANQX", "id": 70904, "logo": null, "date": "2025-10-02T11:50:00+02:00", "start": "11:50", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70904-symbolic-learning-and-rule-extraction-with-sole-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZLANQX/", "title": "Symbolic Learning and Rule Extraction with Sole.jl", "subtitle": "", "track": "Machine learning", "type": "Talk", "language": "en", "abstract": "**Symbolic learning** is a branch of **machine learning** that studies algorithms for building **symbolic models**, classifiers that can be translated to **logical rules**. Hence, differently from neural networks and other statistical models, symbolic models are more easily readable and interpretable. Common examples include **decision trees** and their ensemble counterpart, **random forests**.\r\n\r\n**Sole.jl** is a Julia package for symbolic learning and rule extraction, aimed at guiding the user throughout the whole process, from the initial **data** to learning a symbolic **model**, to the extraction and manipulation of logical **rules** from such model.\r\n\r\nIt is also the entry point to the **SOLE** framework (which stands for **S**ymb**O**lic **LE**arning), an open-source project fully developed in Julia, serving as an interface to both its main packages, such as **SoleData.jl**, **SoleModels.jl**, **SolePostHoc.jl**, as well as useful packages developed by the learning community, such as **DecisionTree.jl**, **ModalDecisionTrees.jl**, **ModalAssociationRules.jl**, **MLJ.jl**, **XGBoost.jl**, and so on.\r\n\r\nFor instance, given a dataset, one can use **Sole.jl** to fit a **decision tree** model, provided by the **DecisionTree.jl** package, and then extract and eventually manipulate rules from such model through the **SolePostHoc.jl** package. Furthermore, if working with more non-tabular data like images or time-series, one can leverage the **SoleData.jl** package to interpret the dataset as a set of logical interpretations of a more-than propositional logic (e.g., interval or rectangle modal logic) and use a learning package compatible with such logic, such as **ModalDecisionTrees.jl** or **ModalAssociationRules.jl**.\r\n\r\nIn this presentation, we will have a look at how **Sole.jl** works through a hands-on tutorial, emphatizing on its comprehensiveness and user-friendliness. This will also allow us to introduce two newcomers to the **SOLE**  ecosystem: **ModalAssociationRules.jl**, a package for mining association rules between instances, and **SolePostHoc.jl**, a package to extract, interpret and simplify sets of rules starting from a symbolic model.", "description": "**Sole.jl** stems from the idea of simplifying the use of machine learning models for symbolic data analysis, a set of techniques providing logical description of data, valued for its interpretability and explainability. It is designed for both newcomers who are unfamiliar with machine learning concepts, as well as experts in the field who will find in **Sole.jl** a secure, reliable, and fully customizable system.\r\n\r\nAlso designed for extended use, **Sole.jl** offers full customization and supports extended functionalities such as:\r\n- **checking** the parameters submitted by the user;\r\n- handling basic **preprocessing** tasks, such as dataset partitioning and cross-validation;\r\n- performing **training**, also allowing for parameters tuning, and **testing**;\r\n- printing analysis **results** also allowing to extract, simplify and manipulate rules and models leveraging the **SolePostHoc.jl** package.\r\n\r\nThe **Sole.jl** engine can be operated with just a single line of code. For instance, we can model the theory underlying a dataset with different strategies and compare them using:\r\n\r\n```julia\r\nmodelset = symbolic_analysis(X; model=(type=:randomforest, :modaldecisiontree))\r\n```\r\n\r\nwhere `X` denotes the dataset under analysis and `model` specifies the machine learning model to be employed. For example, the former could consist of a set of body movement recordings performed by various athletes, the latter can be one of the many models provided by the **SoLe** (**S**ymb**o**lic **Le**arning; also, sun in italian) framework, for instance, through the **SoleModels.jl**, **ModalDecisionTrees.jl**, or **ModalAssociationRules.jl** packages, as well as other widely used packages from the learning community such as **MLJ.jl**, **DecisionTree.jl**, **XGBoost.jl**, and so on.\r\n\r\nAs an example, we will focus on **ModalAssociationRules.jl**, one of the latest novelties in the Julia ecosystem.\r\n\r\n**ModalAssociationRules.jl** is a package for mining and analysing association rules from logisets (i.e., sets of logical interpretations onto which formulas can be checked).\r\n\r\n**Association rules** are co-occurrence relations between **patterns**, that is, features or events that frequently occur together in the data. For example, given a dataset encoding arm movements, an association rule shared by many movements might be:\r\n\r\n```julia\r\nright_hand_moves_right AND <A>right_hand_moves_left THEN <B>left_hand_moves_up\r\n```\r\nwhich can be read as \"when the right hand moves to the right in a certain range of time and for a certain distance, and, after that movement, the same hand moves to the left, then it must be that the whole movement started with the left hand moving up\".\r\n\r\nThe `Miner` structure serves as the core for the whole package, wrapping everything necessary to perform mining: \r\n\r\n- the **alphabet** within which to count co-occurrences;\r\n- the **logiset** representation of a dataset;\r\n- a **mining** algorithm;\r\n- a vector of **measures** to compute for each rule;\r\n- a set of **policies** to reduce the search space.\r\n\r\nOne of the strengths of the **ModalAssociationRules.jl** package is the support for mining **temporal relations**, providing the implementation for two well-known algorithms generalised to handle **modal logics** and, thus, potentially deal with temporal, spatial, spatio-temporal, and other kinds of unstructured data.\r\n\r\nFinally, **SolePostHoc.jl** provides **knowledge extraction** algorithms through a uniform interface, allowing for the comparison of different post-hoc interpretation methods while maintaining a coherent and intuitive user experience:\r\n\r\n```julia\r\nstruct ALGORITHMNAME <: RuleExtractor end\r\nmodalextractrules(:ALGORITHMNAME, model, args...)\r\n```\r\n\r\n**SolePostHoc.jl** integrates a wide range of algorithms for knowledge extraction, including:\r\n\r\n- **Surrogate Trees**, algorithms that approximate complex models such as neural networks or random forests with more interpretable decision trees;\r\n- **Knowledge Distillation**, techniques for transferring knowledge from complex models to simpler and more transparent ones;\r\n- **Rule Extraction**, methods for deriving clear and understandable logical rules from any machine learning model.\r\n\r\nGoing back to our example, consider the following theory, consisting of a list of logical rules, obtained through a Decision Tree provided by the **DecisionTree.jl** package:\r\n\r\n```\r\nIF right_hand_moves_right AND <A>right_hand_moves_left THEN <B>left_hand_moves_up;\r\n\r\nOTHERWISE, IF right_hand_moves_right AND (not <A>right_hand_moves_left) THEN <B>left_hand_moves_up.\r\n```\r\n\r\nwe can leverage **SolePostHoc.jl** to obtain the following new theory, which is more succinct and equally expressive:\r\n\r\n```\r\nIF right_hand_moves_right THEN <B>left_hand_moves_up.\r\n```\r\n\r\nLooking into the future, the **development roadmap** for **Sole.jl** includes:\r\n- features **pre-selection**;\r\n- support to **load** and **store** models;\r\n- **formatted** outputs, delivering results in both **PDF** and **LaTeX** form;\r\n- a **GUI** to further enhance the **user experience**.", "recording_license": "", "do_not_record": false, "persons": [{"code": "UDB889", "name": "Alberto Paparella", "avatar": "https://pretalx.com/media/avatars/UDB889_Ui4U2wy.webp", "biography": "Hello everyone! My name is Alberto Paparella, and I am currently a PhD student in Mathematics at the University of Ferrara. My main interests are Mathematical Logic, specifically Many-Valued and Modal Logics, and Machine Learning. In the last few years, I have been working with the Applied Computational Logic and Artificial Intelligence Laboratory on the SOLE framework for Symbolic Learning in Julia, where my main contributions have been a sub-module for the SoleLogics.jl core package to work with Many-Valued Logics and a package for satisfiability and authomated theorem proving for Many-Valued Multi-Modal Logic based on analytic tableau technique, namely SoleReasoners.jl.", "public_name": "Alberto Paparella", "guid": "5266555b-99fb-53cb-9f82-1ef91535454d", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/UDB889/"}, {"code": "QPMGYE", "name": "Mauro", "avatar": "https://pretalx.com/media/avatars/QPMGYE_BKESh18.webp", "biography": "Hi, I am Mauro Milella, and I am currently a master student at the University of Ferrara. In the last few years, I have been collaborating with the Applied Computational Logic and AI (ACLAI) Laboratory, where we have been developing the Sole framework for symbolic learning entirely in Julia.", "public_name": "Mauro", "guid": "3e1a40f6-9170-58af-b3ec-7a5061c8d9c4", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/QPMGYE/"}, {"code": "MYUTRS", "name": "Marco Perrotta", "avatar": "https://pretalx.com/media/avatars/MYUTRS_FqrYiCb.webp", "biography": "Hi, my name is Marco Perrotta. I'm a computer science student at the University of Ferrara, where I also work as a collaborator at the Applied Computational Logic and Artificial Intelligence Lab. My main interest is how technology can be used to understand and study language.", "public_name": "Marco Perrotta", "guid": "1c8ca06d-b654-5d01-93bc-d5c703b91406", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/MYUTRS/"}, {"code": "HQH3AA", "name": "Riccardo Pasini", "avatar": null, "biography": "Hi, I am Riccardo Pasini, and I am currently a student at the University of Ferrara. In the last few years, I have been collaborating with the Applied Computational Logic and AI (ACLAI) Laboratory, where we have been developing the Sole framework for symbolic learning entirely in Julia.", "public_name": "Riccardo Pasini", "guid": "04188441-d731-5cd0-9ae8-e18281199f62", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HQH3AA/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZLANQX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZLANQX/", "attachments": []}, {"guid": "fc6ecfb7-fef6-5d4a-88ec-c7f2aeca0801", "code": "3YRRM3", "id": 70631, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/3YRRM3/BetaML_log_ykUqzMw.png", "date": "2025-10-02T12:20:00+02:00", "start": "12:20", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70631-new-features-of-the-beta-machine-learning-toolkit-betaml-missing-value-imputation-autoencoders-and-variable-importance-metrics", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/3YRRM3/", "title": "New Features of the Beta Machine Learning Toolkit (BetaML): Missing Value Imputation, Autoencoders, and Variable Importance Metrics", "subtitle": "", "track": "Machine learning", "type": "Lightning talk", "language": "en", "abstract": "[BetaML.jl](https://github.com/sylvaticus/BetaML.jl/) is a lightweight, pure-Julia machine learning library with a scikit-learn-like, consistent API, designed for usability over complexity. It supports decision trees, neural networks, clustering algorithms, and essential ML workflows, with heuristics that simplify usage and a one-parameter autotuning system. This talk highlights its new features: missing value imputation, non-linear dimensionality reduction via autoencoders, and variable importance metrics.", "description": "In the contemporary landscape of machine learning (ML), there is an increasing demand for frameworks that prioritize accessibility and usability, particularly in educational settings, rapid prototyping, and research workflows. [BetaML.jl](https://github.com/sylvaticus/BetaML.jl) responds to this demand with a lightweight, pure Julia library that offers a consistent and intuitive interface, inspired by the design principles of scikit-learn. The library emphasizes ease of use and low entry barriers, while still supporting a rich set of ML models and utilities tailored to both novice and advanced users.\r\n\r\nAt the core of BetaML.jl is a scikit-learn-like API centered on a small set of consistent functions: model instantiation through `Model()`, training with `fit!`, prediction using `predict`, and, where applicable, reconstruction of input data via `inverse_predict`. This uniform interface fosters code clarity and supports a shallow learning curve, particularly for users transitioning from Python-based ML workflows. Importantly, the library adopts a usability-first philosophy: hyperparameters come with meaningful defaults, many are automatically inferred through heuristic strategies, and a unified one-parameter autotuning mechanism enables efficient model calibration with minimal user intervention.\r\n\r\nBetaML.jl provides implementations for both supervised and unsupervised learning. In the domain of supervised learning, the library includes decision trees and random forests\u2014offering interpretable, non-parametric models suitable for both classification and regression\u2014as well as customizable feedforward neural networks. These are implemented in pure Julia and are designed to handle small to medium-sized datasets efficiently. For unsupervised learning, BetaML.jl includes clustering algorithms such as K-Means and K-Medoids for hard partitioning, alongside Gaussian Mixture Models (GMMs) for probabilistic soft clustering and density estimation. All models conform to the same API paradigm and can be easily integrated into broader workflows.\r\n\r\nIn addition to the core models, BetaML.jl incorporates an extensive suite of utilities that facilitate the end-to-end machine learning pipeline. These include data preprocessing tools such as scalers and encoders, sampling methods for validation (including k-fold cross-validation), and a variety of loss functions and performance metrics. These components are designed to work seamlessly with the predictive models, enhancing reproducibility and reducing boilerplate code.\r\n\r\nThis presentation will also highlight a set of newly introduced features that significantly extend the library\u2019s functionality. First, **missing value imputation** is now natively supported through a suite of imputation models. These include simple heuristics such as mean and median imputation, as well as more sophisticated approaches based on Gaussian mixture models, random forests, or general-purpose regressors. The imputation framework supports iterative estimation, multiple imputations, and integration with external models, offering robustness in the face of incomplete data.\r\n\r\nSecond, BetaML.jl introduces **autoencoders** as part of its non-linear dimensionality reduction toolkit. Autoencoders are implemented as deep neural networks capable of learning compressed, information-preserving representations of high-dimensional data. This functionality is particularly relevant for feature extraction, denoising, and visualization of complex data structures. In addition, a PCA-based linear encoder is available, providing users with both linear and non-linear dimensionality reduction options within the same framework.\r\n\r\nA third major addition is the provision of **variable importance metrics**, aimed at improving model interpretability. The new `FeatureRanker` utility offers both mean decrease in accuracy and variance decomposition (SOBOL index) importance scoring, allowing users to quantify the contribution of each input variable to model performance. This feature supports multiple ranking strategies, including permutation-based scoring, \u201cpermute and relearn\u201d, or\u2014if supported by the estimator\u2014\u201cfit once and omit in prediction.\u201d It is compatible with any estimator, not necessarily from BetaML, that provides a `fit`/`predict` API.\r\n\r\nTo conclude the session, the talk will include a demonstration of a complete ML pipeline using BetaML.jl. This will involve loading and preprocessing a dataset, training a model with automatic hyperparameter tuning, applying autoencoders for dimensionality reduction, and interpreting model predictions through variable importance analysis. The goal is to provide attendees with a practical understanding of how BetaML.jl can be used to construct efficient and reproducible ML workflows entirely within the Julia ecosystem.\r\n\r\nIn sum, BetaML.jl offers a coherent, user-oriented design that balances simplicity and power. It serves both as an educational platform and a research-grade tool for applied machine learning in Julia, with particular attention to model transparency, ease of deployment, and extensibility.", "recording_license": "", "do_not_record": false, "persons": [{"code": "BWZXEB", "name": "Antonello Lobianco", "avatar": "https://pretalx.com/media/avatars/BWZXEB_Xh8lx7s.webp", "biography": "[Antonello Lobianco](https://github.com/sylvaticus), PhD, is a research engineer employed by a French grande \u00e9cole (polytechnic university). He works on the biophysical and economic modeling of the forest sector and is responsible for the lab models portfolio. He programs in C++, Perl, PHP, Visual Basic, Python, and Julia. He teaches environmental and forest economics at the undergraduate and graduate levels and modeling at the PhD level. For a several years, Antonello has been following the development of Julia, as it fits his modeling needs. He is the author of a few Julia packages, particularly on data analysis and machine learning", "public_name": "Antonello Lobianco", "guid": "f31919a4-cf23-56af-bf61-59b101c6f07c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/BWZXEB/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/3YRRM3/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/3YRRM3/", "attachments": []}, {"guid": "d8a14f43-ee3c-5144-9242-f18bc63f4f91", "code": "RYGM79", "id": 69597, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/RYGM79/rl-logo_Qk6soS7.png", "date": "2025-10-02T12:30:00+02:00", "start": "12:30", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-69597-extending-f-lux-jl-s-recurrent-neural-network-offerings-with-recurrentlayers-jl-and-luxrecurrentlayers-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/RYGM79/", "title": "Extending F/Lux.jl\u2019s Recurrent Neural Network Offerings with RecurrentLayers.jl and LuxRecurrentLayers.jl", "subtitle": "", "track": "Machine learning", "type": "Lightning talk", "language": "en", "abstract": "Recurrent neural networks (RNNs) are deep learning (DL) models for sequential data, with many variants developed over the last two decades. However, core DL libraries typically offer only a limited number of cell implementations. In this talk, I introduce [RecurrentLayers.jl](https://github.com/MartinuzziFrancesco/RecurrentLayers.jl) and [LuxRecurrentLayers.jl](https://github.com/MartinuzziFrancesco/LuxRecurrentLayers.jl), two libraries that extend Flux and Lux with over 30 additional RNN cells each, enabling broader experimentation and research.", "description": "Recurrent neural networks (RNNs) are widely used machine learning models (ML) for modeling sequential data. By maintaining an internal state through recurrent connections, RNNs can keep a dynamical memory of past time steps and capture temporal dependencies. Over the years, many RNN variants have been proposed to improve learning dynamics, stability, and expressiveness. As a result, the landscape of recurrent models has grown increasingly rich and diverse.\r\n\r\nHowever, these models are often scattered across unique implementations, in different frameworks and one-off repositories, making comparison and experimentation difficult. The advent of the transformers era further contributed to relegating RNNs into a secondary role, leading to reduced support and limited availability of diverse architectures.\r\n\r\nAs a solution, I introduce [RecurrentLayers.jl](https://github.com/MartinuzziFrancesco/RecurrentLayers.jl) and [LuxRecurrentLayers.jl](https://github.com/MartinuzziFrancesco/LuxRecurrentLayers.jl), two libraries that expand the RNN offerings of Flux.jl and Lux.jl. The libraries provide each over 30 recurrent cells and modular components for building custom architectures. Core RNN enhancements, such as multiplicative integration and independent recurrence, are implemented in a reusable way.\r\n\r\nIn this talk, I will present an overview of the packages, demonstrate how they integrate with the existing deep learning ecosystem in Julia, and highlight the range of features they offer for research and experimentation. Finally, I will showcase how the libraries facilitate easy cross-pollination of ideas from different research papers, enabling the rapid development of novel and customized RNN architectures.", "recording_license": "", "do_not_record": false, "persons": [{"code": "NGLZHC", "name": "Francesco Martinuzzi", "avatar": "https://pretalx.com/media/avatars/NGLZHC_qn3Wicu.webp", "biography": "I\u2019m a postdoc at the Max Planck Institute for the Physics of Complex Systems in Dresden, with research interests in machine learning, chaos, and nonlinear time series. My current work focuses on using ML to forecast chaotic dynamics.\r\n\r\nI\u2019ve been a Julia user since 2019 and have contributed to the ecosystem with packages like [ReservoirComputing.jl](https://github.com/SciML/ReservoirComputing.jl) and [CellularAutomata.jl](https://github.com/MartinuzziFrancesco/CellularAutomata.jl).\r\n\r\nYou can find out more about my work on [my website](https://martinuzzifrancesco.github.io/).", "public_name": "Francesco Martinuzzi", "guid": "a8b2dd8d-3e80-5c46-90a5-64b401e7b631", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/NGLZHC/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RYGM79/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RYGM79/", "attachments": []}, {"guid": "fe064080-3e75-5060-872d-15957d2497c3", "code": "RZD3UX", "id": 72232, "logo": null, "date": "2025-10-02T12:40:00+02:00", "start": "12:40", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-72232-user-friendly-inference-with-rxinfer-s-projectedto-constraints", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/RZD3UX/", "title": "User-friendly Inference with RxInfer's ProjectedTo Constraints", "subtitle": "", "track": "Machine learning", "type": "Lightning talk", "language": "en", "abstract": "Traditional message passing-based Bayesian inference algorithms are fast but often struggle with real-world models that lack closed-form update rules, such as those with non-conjugate terms or any external code integrated into RxInfer: differential equation solvers, root finding, or even code outside of Julia. The Bayesian inference package RxInfer (from v3) introduces a powerful projection mechanism that overcomes this limitation, enabling modelers to maintain the speed of message passing while handling complex model components.\r\n\r\nThe core idea is simple: you define your joint distribution in Julia, then annotate challenging factors with `ProjectedTo(<family>)` constraints. RxInfer then automatically derives the necessary inference approximations and delivers streaming estimates of posterior distributions. This talk will demonstrate this workflow through two compact examples: a hierarchical Gaussian filter for choice data and integrating an ordinary differential equation (ODE) solver for an SIR epidemic model. Attendees will receive runnable notebooks and a clear recipe for applying reactive variational inference to their own non-conjugate or simulator-based models.", "description": "Message passing offers very fast Bayesian inference calculations, but its power has historically been limited to dealing with conjugate priors and likelihoods. Real-world models rarely fit neatly into this box. In contrast, RxInfer-v3 solves this issue with its new `@constraints q(x)::ProjectedTo(<family>)` directive. You simply tag an edge (a random variable in your model, `x` in this case) that is involved around a ``stubborn'' factor (a probabilistic dependency between your random variables) with the desired exponential-family form, and the library automatically re-projects every update, keeping inference efficient.\r\n\r\nWe will build a concrete example live on stage: a non-conjugate hierarchical Gaussian filter for choice data, all in under 70 lines of Julia code. We will show how to declare just two `ProjectedTo` constraints, hit ``infer,'' and watch posteriors stream without ever needing to write manual update equations.\r\n\r\nNext, we will demonstrate how the very same trick can wrap an external ODE solver as a ``black-box'' factor. This powerful capability means you can seamlessly integrate differential equation models (or any C, Fortran, or Python routine) directly into your graph while maintaining reactivity.\r\n\r\nAttendees will leave with copy-paste notebooks and a clear mental model for fearless, rule-free probabilistic modeling.\r\n\r\n### Key Takeaways for Attendees:\r\n\r\n- How the new `@constraints ProjectedTo` replaces tedious manual rule writing.\r\n- Practical patterns for plugging ODEs, simulators, or any external C/Python routine into a factor graph.\r\n- A self-contained notebook, a mental recipe for fearless modelling, and zero new mathematical derivations to memorize", "recording_license": "", "do_not_record": false, "persons": [{"code": "F87RD7", "name": "Dmitry Bagaev", "avatar": "https://pretalx.com/media/avatars/F87RD7_AUZwJIn.webp", "biography": null, "public_name": "Dmitry Bagaev", "guid": "cd9c7313-928b-5655-8a22-4fcde34e5384", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/F87RD7/"}, {"code": "FEWRRE", "name": "Mykola Lukashchuk", "avatar": "https://pretalx.com/media/avatars/FEWRRE_m9CAN0P.webp", "biography": "Mykola Lukashchuk is a PhD candidate in Electrical Engineering at Eindhoven University of Technology, focusing on probabilistic inference and efficient Bayesian computation. His research develops flexible computational engines that trade precision for efficiency through message-passing algorithms and Riemannian manifold representations.\r\n\r\nHe holds dual Master's degrees in Statistics (Kyiv University) and Computer Science (Instituto Polit\u00e9cnico Nacional). His work contributes to the RxInfer ecosystem with particular focus on ExponentialFamilyManifolds.jl and efficient Bayesian inference methods.", "public_name": "Mykola Lukashchuk", "guid": "3511aaad-3b95-53b6-8cc6-cf35805c6fa9", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/FEWRRE/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RZD3UX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RZD3UX/", "attachments": []}, {"guid": "57f7a000-31f6-5a18-8e07-31dbbe1116ca", "code": "RMNNTW", "id": 81311, "logo": null, "date": "2025-10-02T12:50:00+02:00", "start": "12:50", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-81311-sponsor-talk-pumas-ai", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/RMNNTW/", "title": "Sponsor talk - Pumas AI", "subtitle": "", "track": "General", "type": "Sponsor talk", "language": "en", "abstract": "Sponsor talk by Pumas AI", "description": "-", "recording_license": "", "do_not_record": false, "persons": [{"code": "GDCCSG", "name": "Andreas Noack", "avatar": null, "biography": null, "public_name": "Andreas Noack", "guid": "6b87c2f1-766f-52e7-8d90-ea87227c264e", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/GDCCSG/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RMNNTW/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RMNNTW/", "attachments": []}, {"guid": "96fb2e23-dede-537d-aa11-a51c96185b34", "code": "SGZGMA", "id": 79777, "logo": null, "date": "2025-10-02T14:30:00+02:00", "start": "14:30", "duration": "01:00", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79777-kernels-without-borders-parallel-programming-with-kernelabstractions-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/SGZGMA/", "title": "Kernels without borders: Parallel programming with KernelAbstractions.jl", "subtitle": "", "track": "GPU & performance", "type": "Plenary", "language": "en", "abstract": "Modern computing relies on parallelism, from GPUs accelerating AI workloads to multi-core CPUs in every laptop. But writing code that harnesses this power across different hardware remains challenging. In this talk, we'll explore how KernelAbstractions.jl brings GPU-style programming to Julia, allowing you to write parallel kernels once and run them anywhere.", "description": "Tim Besard is a software engineer at JuliaHub, where he leads GPU support and development for the Julia programming language. He holds a Ph.D. in computer science engineering from Ghent University, Belgium, and has been a key contributor to Julia's GPU ecosystem since 2014. Tim maintains several foundational GPU packages including CUDA.jl, GPUArrays.jl, GPUCompiler.jl, and LLVM.jl, which together form the backbone of GPU computing in Julia.", "recording_license": "", "do_not_record": false, "persons": [{"code": "9YQMAK", "name": "Tim Besard", "avatar": "https://pretalx.com/media/avatars/9YQMAK_BNCDK1H.webp", "biography": "Tim Besard is a software engineer at JuliaHub, where he leads GPU support and development for the Julia programming language. He holds a Ph.D. in computer science engineering from Ghent University, Belgium, and has been a key contributor to Julia's GPU ecosystem since 2014. Tim maintains several foundational GPU packages including CUDA.jl, GPUArrays.jl, GPUCompiler.jl, and LLVM.jl, which together form the backbone of GPU computing in Julia.", "public_name": "Tim Besard", "guid": "40c37e38-af42-51ce-9f6a-cbf1d84adf2a", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/9YQMAK/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/SGZGMA/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/SGZGMA/", "attachments": []}, {"guid": "a933402c-ecaf-552f-96cb-1e02c0cbe90e", "code": "HZUFYF", "id": 70614, "logo": null, "date": "2025-10-02T15:30:00+02:00", "start": "15:30", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70614-solving-optimal-control-problems-on-gpu", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/HZUFYF/", "title": "Solving optimal control problems on GPU", "subtitle": "", "track": "GPU & performance", "type": "Lightning talk", "language": "en", "abstract": "Optimal control problems of ordinary differential equations are challenging numerical problems to solve with applications in many fields, including a current strong trend in quantum control. The aim of this talk is to present the ongoing effort made in the framework of control-toolbox.org with the Julia package OptimalControl.jl, within a collaboration involving Inria, CNRS and Argonne. The focus will be on so-called direct methods that discretise the control problem into a large sparse nonlinear mathematical program. The numerical solution then heavily relies on sparse automatic differentiation (AD) and sparse numerical linear algebra for optimisation solvers. This includes remarkable progress in Julia with such packages as Enzyme.jl, ADNLPModels.jl, ExaModels.jl or MadNLP.jl. The last two packages, in particular, allow to take advantage of the natural massive SIMD parallelism of the discretised problem to provide efficient solving on GPU's.", "description": "The specific structure of the discretised math program problem raises challenging issues in terms of sparsity detection for AD, numerical linear algebra and optimisation. Going to GPU also implies strong modelling and discretisation choices to achieve efficient solving for a large number of variables and constraints. These topics are at the heart of a beginning collaboration involving people from Inria (McTAO team) and Argonne (notably A. Montoison), and also benefits from an existing JLESC project between the two institutions (\"Shared Infrastructure for Source Transformation Automatic Differentiation\").", "recording_license": "", "do_not_record": false, "persons": [{"code": "HEYELZ", "name": "Jean-Baptiste Caillau", "avatar": "https://pretalx.com/media/avatars/HEYELZ_Cw98jAD.webp", "biography": "Professor of applied math at Universit\u00e9 C\u00f4te d\u2019Azur, CNRS, Inria, LJAD\r\n\r\nScientific interests - Optimisation and control: geometry, algorithms, applications\r\n\r\nhttps://caillau.perso.math.cnrs.fr", "public_name": "Jean-Baptiste Caillau", "guid": "26a442d5-4946-5706-8745-fb1073f640c7", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HEYELZ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/HZUFYF/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/HZUFYF/", "attachments": []}, {"guid": "a7823410-6100-566c-aafd-f94352433e75", "code": "VYEUUV", "id": 70824, "logo": null, "date": "2025-10-02T15:40:00+02:00", "start": "15:40", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70824-reactant-jl-optimize-julia-functions-with-mlir-and-xla-for-high-performance-execution-on-cpu-gpu-tpu-and-more", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/VYEUUV/", "title": "Reactant.jl - Optimize Julia Functions With MLIR and XLA for High-Performance Execution on CPU, GPU, TPU and more.", "subtitle": "", "track": "GPU & performance", "type": "Lightning talk", "language": "en", "abstract": "Reactant.jl is a package that takes a Julia function operating on arrays, and compiles it into the MLIR representation. It can then run fancy optimizations on top of it, including using EnzymeMLIR for automatic differentiation, and create relevant executables for CPU/GPU/TPU via XLA.", "description": "Reactant.jl is a package that takes a Julia function operating on arrays, and compiles it into the MLIR representation. It can then run fancy optimizations on top of it, including using EnzymeMLIR for automatic differentiation, and create relevant executables for CPU/GPU/TPU via XLA. \r\n\r\nIn this talk, we go over the advantages of such a system for a variety of tensor related workloads where having tensor aware optimization becomes beneficial. We describe the inner workings of the package and its current limitations. We also demonstrate the current raising capabilities of Reactant, enabling running CUDA.jl kernels on other accelerators such as CPU, TPU or AMD GPUs. Finally, we highlight successful integrations in Julia packages such as Lux.jl and Tenet.jl.", "recording_license": "", "do_not_record": false, "persons": [{"code": "9QBUME", "name": "Sergio S\u00e1nchez Ram\u00edrez", "avatar": "https://pretalx.com/media/avatars/9QBUME_B2JWIIp.webp", "biography": "PhD student at Barcelona Supercomputing Center (BSC-CNS)", "public_name": "Sergio S\u00e1nchez Ram\u00edrez", "guid": "07debff1-9cb7-5a7f-99cb-4e30da8c372f", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/9QBUME/"}, {"code": "8A7MJS", "name": "Paul Berg", "avatar": "https://pretalx.com/media/avatars/8A7MJS_VHgAHTO.webp", "biography": "Interested about Julia and developer tools.", "public_name": "Paul Berg", "guid": "8bbf98cf-e02b-5405-9e88-adddddfd6ac6", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/8A7MJS/"}, {"code": "ZMNTFE", "name": "Ga\u00ebtan LOUNES", "avatar": null, "biography": "3nd year PhD student at IRISA (Institute for Research in Computer Science and Random Systems) interested in compilers, MLIR and FPGA.", "public_name": "Ga\u00ebtan LOUNES", "guid": "6f8369a4-5f7d-5845-8f5d-0aadfa9cebf1", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/ZMNTFE/"}, {"code": "HDYDC9", "name": "Jules", "avatar": null, "biography": null, "public_name": "Jules", "guid": "673ffc7b-2c12-5eb5-b808-8458014f587e", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HDYDC9/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VYEUUV/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VYEUUV/", "attachments": []}, {"guid": "4cefa816-5e52-5765-bcf8-18f7e576d36f", "code": "9L97JX", "id": 67063, "logo": null, "date": "2025-10-02T15:50:00+02:00", "start": "15:50", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-67063-vulkan-jl-cross-platform-graphics-on-the-gpu", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/9L97JX/", "title": "Vulkan.jl: cross-platform graphics on the GPU", "subtitle": "", "track": "GPU & performance", "type": "Lightning talk", "language": "en", "abstract": "We present an idiomatic Julia interface to Vulkan, a high-performance cross-platform graphics API for GPUs. We will briefly introduce Vulkan and outline reasons for using it and situations where it is a solid choice, then present Vulkan.jl as the most convenient way known to date to use Vulkan from Julia. We will quickly go through its main features, all about automating everything you would do by hand and in a better way. As an extremely large and complex specification, Vulkan is hard to manually interface with while guaranteeing correctness of execution, which is precisely what Vulkan.jl aims to solve. We will end with a few example applications that make use of this package.", "description": "[Vulkan.jl](https://github.com/JuliaGPU/Vulkan.jl) is a lightweight wrapper around the\u00a0[Vulkan](https://www.vulkan.org/)\u00a0graphics and compute library. It exposes abstractions over the underlying C interface, primarily geared towards developers looking for a more natural way to work with Vulkan with minimal overhead. It builds upon the core API provided by\u00a0[VulkanCore.jl](https://github.com/JuliaGPU/VulkanCore.jl/), but because Vulkan is originally a C specification and available via C interop, interfacing with it requires some knowledge to be used correctly from Julia. This package acts as an abstraction layer, so that one doesn't need to deal with C interoperability details, while still retaining full functionality and maximum performance.\r\n\r\nMany of the patterns encountered in the Vulkan specification are very common; thus, we were able to understand and automate most of them with advanced implementations that attempt to provide the most robust and performant way to lift a C API to a more idiomatic Julia API, making use of convenience features available in the Julia language.", "recording_license": "", "do_not_record": false, "persons": [{"code": "9N8YYQ", "name": "C\u00e9dric Belmant", "avatar": "https://pretalx.com/media/avatars/9N8YYQ_X6pwipt.webp", "biography": "C\u00e9dric is an applied mathematician and a programmer, enjoying developing tools for graphics applications on his free time.\r\nHe currently works at JuliaHub as part of the Compiler team.", "public_name": "C\u00e9dric Belmant", "guid": "972d651c-8c70-578c-9d8d-0a11aae7b185", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/9N8YYQ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/9L97JX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/9L97JX/", "attachments": [{"title": "Slides", "url": "/media/juliacon-local-paris-2025/submissions/9L97JX/resources/_daaSS5n.pdf", "type": "related"}]}, {"guid": "d39dfb65-b1d4-5e73-843d-ac8a7bfc166b", "code": "GDZY7X", "id": 72290, "logo": null, "date": "2025-10-02T16:00:00+02:00", "start": "16:00", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-72290-why-are-sum-functions-extremely-hard-to-get-right", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/GDZY7X/", "title": "Why are sum functions extremely hard to get right?", "subtitle": "", "track": "GPU & performance", "type": "Talk", "language": "en", "abstract": "Writing a loop to add up all the elements in an array is often featured in tutorials and training materials \u2014\u00a0it's just five lines of code, and can be remarkably generic and performant.  Yet there are a surprising number of challenges in determining exactly how such a function should operate while ensuring consistent generic behaviors, numeric accuracy, high performance, and an extensible interface.\r\n\r\nThis talk will describe these challenges and some of their solutions, looking at both the specific designs of Julia's builtin reduction machinery (and how it may change) as well as general learnings about best practices when designing such highly generic, performance-critical, and extensible interfaces.", "description": "Base Julia's reduction machinery is largely the same as it was originally implemented over a decade ago, and it implements highly generic behaviors that apply across many different types of containers. It has many orthogonal features and switches.  This includes:\r\n\r\n* The three functions with differing associativities: `foldl` vs. `reduce` vs. `foldr`\r\n* Whether an `init` was provided or not \u2014\u00a0and if so, how that `init` should be used\r\n* Whether `dims` was provided or not \u2014\u00a0and if so, if the first dimension is included\r\n* If the container is iterable or array-like \u2014 and if so, whether its indices are linear or cartesian\r\n* If the container has zero, one, or more elements\r\n* If there's a known reducing and/or mapping function being applied \u2014\u00a0and if it's ok to speculatively evaluate or memoize calls of the passed functions\r\n\r\nIn short: it's complicated!  It's also highly performance sensitive.  Everyone expects the `sum` function to be both fast and accurate,\u00a0and while there are cases where those two criteria align, they can also be at adds with one another.\r\n\r\nJulia's current documentation hasn't fully specified how many of these features should work, but a number of devs have been working over the past few years to more rigorously define what the behaviors should be.  As consensuses around the design fundamentals have grown, it's become more clear what incremental changes are possible and how a completely reimagined implementation could work.  Building out that reimagined implementation has not been trivial, and it brings together many learnings of best practice generic Julia programming.", "recording_license": "", "do_not_record": false, "persons": [{"code": "VUPRBT", "name": "Matt Bauman", "avatar": "https://pretalx.com/media/avatars/VUPRBT_JtQKHHN.webp", "biography": "Matt has been a part of the Julia community for over a decade and is the Director of Sales Engineering at JuliaHub.", "public_name": "Matt Bauman", "guid": "704172c4-42c5-5593-b9e6-151ec3abd92e", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/VUPRBT/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/GDZY7X/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/GDZY7X/", "attachments": []}, {"guid": "9b2f1a48-5209-5b76-b26a-214e776b5054", "code": "NYHBSL", "id": 70635, "logo": null, "date": "2025-10-02T17:00:00+02:00", "start": "17:00", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70635-uncertaintyquantification-jl-efficient-uncertainty-propagation-powered-by-julia", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/NYHBSL/", "title": "UncertaintyQuantification.jl: Efficient uncertainty propagation powered by Julia", "subtitle": "", "track": "Error, derivatives, stability", "type": "Talk", "language": "en", "abstract": "This talk introduces  UncertaintyQuantification.jl , a generalised framework for uncertainty quantification.  The framework has undergone extensive development since its initial release in August of 2020 and now includes a number of numerical algorithms to quantify and propagate uncertainties.  We have previously presented the package at two scientific conferences and now want to share it with the wider Julia community.\r\n\r\nIn this talk we discuss a significant subset of the features currently available. Adequate illustrative numerical examples from various engineering disciplines are presented throughout, to highlight the capabilities of the implemented algorithms.", "description": "The definition of *uncertainty* follows from the absence of *certainty* describing a state of absolute knowledge where everything there is to know about a process is known.  This however is a theoretical and unachievable state, rendering deterministic models insufficient for safe analysis of engineering systems.\r\n\r\nWith *UncertaintyQuantification.jl* we aim to provide a comprehensive and generalised toolbox to deal with uncertainty across various engineering disciplines. \r\n\r\n## Current features:\r\n- Simulation-based reliability analysis\r\n  - (Advanced) Monte Carlo simulation\r\n    - Quasi Monte Carlo\r\n    - Line Sampling\r\n    - Subset Simulation\r\n - Global Sensitivity analysis\r\n   - Sobol indices\r\n - Metamodeling\r\n   - Response Surface Methodology\r\n   - Polynomial Chaos Expansion\r\n - Bayesian Updating\r\n   - Metropolis-Hastings\r\n   - Transitional Markov Chain Monte Carlo\r\n - Third-party solvers\r\n   - Connect to any solver by injecting random samples into source files\r\n   - HPC interfacing with slurm\r\n - Stochastic Dynamics\r\n   - Power Spectral Density Estimation\r\n   - Stochastic Process Generation\r\n - Imprecise probabilites\r\n   - Intervals\r\n   - Probability boxes\r\n\r\n## Links\r\n - **Github:** https://github.com/FriesischScott/UncertaintyQuantification.jl\r\n - **Documentation:** https://friesischscott.github.io/UncertaintyQuantification.jl/dev/", "recording_license": "", "do_not_record": false, "persons": [{"code": "9L3JQF", "name": "Jasper Behrensdorf", "avatar": "https://pretalx.com/media/avatars/9L3JQF_bhNvqj3.webp", "biography": "Postdoctoral researcher at the *Institute for Risk and Reliability* at the *Leibniz University Hannover*, Germany.", "public_name": "Jasper Behrensdorf", "guid": "30c91599-eace-5637-b022-de2ce23a887a", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/9L3JQF/"}, {"code": "GFTSJ3", "name": "Ander Gray", "avatar": null, "biography": "Computational Scientist at the UK Atomic Energy Authority", "public_name": "Ander Gray", "guid": "5ce5be7f-9b46-5487-b04b-97a77b5eef00", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/GFTSJ3/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NYHBSL/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NYHBSL/", "attachments": []}, {"guid": "848110dd-5eb8-557c-b03f-aed224bcfafa", "code": "W8VHYG", "id": 68697, "logo": null, "date": "2025-10-02T17:30:00+02:00", "start": "17:30", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-68697-intervalarithmetic-jl-v1-0-intervals-that-you-can-trust", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/W8VHYG/", "title": "IntervalArithmetic.jl v1.0 - Intervals that you can trust.", "subtitle": "", "track": "Error, derivatives, stability", "type": "Talk", "language": "en", "abstract": "Floating point numbers are beautiful but are limited. If we push them hard enough, computations that use them become inaccurate, and they are only a very thin sample of all real numbers. In contrast, considering intervals on the real line provides strict guarantee and accuracy and allows to draw general conclusions on continuous subsets of the real line. With intervals, a range of rigorous numerical proofs of mathematical facts can be performed.F or example, the existence or absence of solutions to an equation can be established and proven. By defining all required basic arithmetic operations on intervals, they can be used in arbitrary code, extending their benefits to all sort of calculations.\r\n\r\nThe goal of [IntervalArithmetics.jl](https://github.com/JuliaIntervals/IntervalArithmetic.jl) is to bring this method to the Julia ecosystem, without sacrificing Julia's flexibility. To reach this objective, we must face a difficult challenge: how to guarantee that interval computations stay rigorous while seamlessly being performed in codes that are, in general, unaware of the existence of interval arithmetic?\r\n\r\nIn this talk, I will present the strategies that we have implemented for the 1.0 version of the package which allows safe and easy use of rigorous interval computations in Julia, and how it can be applied to root-finding problems.", "description": "Traditionally, computers work with single numbers, represented as floating points. However, this approach does not provide any guarantee, whether on the calculation itself (due to the accumulation of numerical inaccuracies), or on the behavior of the computation in between sampled points. By instead working with intervals, arbitrary floating point computations can be extended to give rigorous bounds on the true result of a computation, which in turn allows drawing meaningful and provably correct conclusions.\r\n\r\nInterval packages in other languages only allow their interval type to interact with other intervals. This grants them full control on the behavior of intervals, at the cost that user code often needs to be rewritten to become compatible with intervals. Thanks to multiple dispatch, in Julia, the situation is different, and intervals can be injected in arbitrary code, often requiring no adaptation from the user. However, when trying to build a rigorous and highly interoperable interval package in Julia, a few major questions arise that are absent for other implentations:\r\n- How should intervals interact with \"normal\" numbers, whose origin is unknown? An inaccurate number could poison a whole computation.\r\n- How should common operations be extended to intervals? Arithmetic operations (e.g. `+`, `/`) have a clear interval counterpart (defined in the corresponding IEEE standard), but boolean ones (e.g. `==`, or `issubset` for multidimensional interval boxes) are subject to interpretation, and require, in general, the use of ternary logic which is likely incompatible with user code.\r\n\r\nIn this talk, I will introduce the general idea and techniques of interval arithmetic, before presenting our solution to those challenges specific to Julia: using a system of flags to store the quality of the interval computation, balancing usability and strictness, and some of the tools we introduced to streamline the experience of exigent users.\r\n\r\nThen, I will showcase how intervals can be used in unaware code, through two successful applications of our work:\r\n- The interoperability of IntervalArithmetic.jl with ForwardDiff.jl, which, thanks to a small package extension, provides guaranteed bounds on the derivative of arbitrary functions.\r\n- The rigorous determination of all roots of a function with IntervalRootFinding.jl, which itself relies on ForwardDiff.jl, and can be used to prove the existence or absence of solutions of an equation in a given domain.", "recording_license": "", "do_not_record": false, "persons": [{"code": "KMMWMD", "name": "Beno\u00eet Richard", "avatar": "https://pretalx.com/media/avatars/KMMWMD_wAt8C88.webp", "biography": "Doctor in theoretical physics, working mostly in simulation and data analysis with experimentalists blowing up tiny things in large facilities.\r\n\r\nWorked on IntervalArithmetic.jl for my master thesis and stuck around.\r\n\r\nMade the questionable life choice of writing a LaTeX engine in julia.\r\n\r\nFirst julia version used: v0.4", "public_name": "Beno\u00eet Richard", "guid": "2eb1a124-3630-547c-bb4d-212df7fed890", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/KMMWMD/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/W8VHYG/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/W8VHYG/", "attachments": []}], "Robert Faure Amphitheater": [{"guid": "422591b2-6c4c-55ea-8c8a-299826244e21", "code": "7ZA99Y", "id": 72320, "logo": null, "date": "2025-10-02T10:00:00+02:00", "start": "10:00", "duration": "01:00", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-72320-how-to-tell-an-interactive-story-with-pluto-lectures-and-dashboards", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/7ZA99Y/", "title": "\ud83c\udf88 How to tell an interactive story with Pluto: lectures and dashboards", "subtitle": "", "track": "Language & teaching", "type": "Tutorial", "language": "en", "abstract": "In this session, the Pluto developers will teach you how to use [Pluto notebooks](https://plutojl.org) for *computational storytelling* to write compelling, interactive articles and dashboards. This tutorial will help you write a lecture for students that's actually more interesting than the latest tiktok trend!", "description": "You already know how Julia can give you amazing results, but how do you share your work? We will explain the core concepts of designing an interactive article or dashboard using Pluto. When is interactivity a cool trick, and when does it actually succeed in explaining a concept in a more intuitive way?\r\n\r\nPluto is a reactive notebook environment for Julia, with interactivity built in. To learn more about Pluto, check out [our website](https://plutojl.org), and to see examples of interactive Pluto notebooks in your browser, browse through our [featured notebooks](https://featured.plutojl.org).\r\n\r\nAt the start of the session, we will share a couple of static notebooks (just the numbers and text), and we will work to make them interactive together. We will share basic concepts like sliders, dynamic text, interactive plots and animation. This is really easy to do, and you will have learned a bunch of interactive techniques at the end of the session!\r\n\r\nWe also go into more complex subjects: combining widgets into dashboard components, layout, typography and coloring, hosting interactive notebooks online, and more! You will get a good understanding of the Pluto ecosystem, and where to look for more information.\r\n\r\nThis tutorial is accessible to all coding levels, beginner to advanced. No prior Pluto knowledge required. Come with a laptop with Julia installed, or bring a friend to work together with.", "recording_license": "", "do_not_record": false, "persons": [{"code": "ECC83Q", "name": "Fons van der Plas", "avatar": "https://pretalx.com/media/avatars/ECC83Q_5CxYWbQ.webp", "biography": "\ud83c\udf3c Main developer of [Pluto.jl](https://plutojl.org/), currently working at TU Eindhoven to develop the course [Bayesian Machine Learning](https://github.com/bmlip/colorized).\r\n\r\nCome talk to me at the conference! I would love to learn more about what you are working on :) If you are teaching using Julia, definitely come talk to me!", "public_name": "Fons van der Plas", "guid": "17bd1580-e1fc-5230-9599-f2cc7f6c0695", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/ECC83Q/"}, {"code": "VAFRRY", "name": "Boshra Ariguib", "avatar": "https://pretalx.com/media/avatars/VAFRRY_lf0lX6G.webp", "biography": "Hi \u2600\ufe0f\u00a0I am a CS Masters student in my final semester at the University of Stuttgart and a big enthusiast of the Pluto.jl Project \ud83c\udf89\r\n\r\nI\u2019m passionate about making coding and science accessible to everyone, regardless of age or background. If you\u2019re working on something along those lines, I\u2019d love to chat! This is my very first conference, so I\u2019m especially excited to meet new people and make friends \ud83d\ude0a", "public_name": "Boshra Ariguib", "guid": "828efe95-6a73-5bd3-ad58-6a17b5f96b75", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/VAFRRY/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7ZA99Y/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7ZA99Y/", "attachments": []}, {"guid": "38f27bb1-1901-5522-bff5-92a568c89ebb", "code": "XSY8DA", "id": 70783, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/XSY8DA/logo_MJUZxb2.jpg", "date": "2025-10-02T11:50:00+02:00", "start": "11:50", "duration": "00:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70783-solving-partial-differential-equations-with-bcube-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/XSY8DA/", "title": "Solving partial differential equations with Bcube.jl", "subtitle": "", "track": "Numerical analysis", "type": "Talk", "language": "en", "abstract": "Bcube.jl is a new open-source Julia package developed at ONERA to solve PDEs on unstructured meshes using finite element and discontinuous Galerkin methods. It enables flexible, efficient, and unified multi-physics simulations. This talk presents the package's philosophy, architecture, and capabilities through diverse examples, and shares insights on using Julia in an industrial research context.", "description": "The Multi-Physics Department for Energetics at ONERA focuses on aerospace applications such as propulsion systems, laminar\u2013turbulent transition, and icing, among others. These phenomena are typically modeled by systems of partial differential equations (PDEs) with diverse mathematical structures. As a result, they are often solved using dedicated solvers with specific numerical methods. However, this fragmented approach complicates the development of strong multi-physics couplings and hinders the mutualization of numerical efforts.\r\nTo address this challenge, we have developed Bcube.jl, an open-source Julia package designed at ONERA for solving PDEs on unstructured meshes. Similar to packages like Gridap.jl and Ferrite.jl, Bcube.jl provides a user-friendly and efficient framework for discretizing PDEs using either the finite element method or the discontinuous Galerkin method. In this talk, we will present the overall philosophy behind Bcube.jl as well as its core architecture.\r\nThe presentation will include illustrative examples across various applications\u2014from the Euler equations around an airfoil to the resolution of PDEs on curved surfaces such as those arising in shallow water models. These examples will demonstrate both the usability and versatility of the library.\r\nWe will conclude with perspectives on the future development of Bcube.jl, particularly in the context of high-performance computing and advanced numerical methods. Finally, we will share reflections on adopting Julia in a research institution like ONERA, which traditionally develops software for the aerospace industry.", "recording_license": "", "do_not_record": false, "persons": [{"code": "BKMQ9V", "name": "Maxime Bouyges", "avatar": "https://pretalx.com/media/avatars/BKMQ9V_xbYrVXv.webp", "biography": "Maxime Bouyges is a Research Engineer in Fluid Dynamics at ONERA's Department of Multi-Physics for Energetics. Commencing his career in 2013, he completed a PhD focusing on stability analysis of complex flows within solid rocket engines. Subsequently, he joined the Multiphase Flow Research Team, contributing to the development of the CEDRE-FILM solver - a parallel 3D surface solver tailored for shallow-water and icing applications. Currently, his work encompasses the advancement of physical models and numerical methods, primarily within the realm of icing phenomena.", "public_name": "Maxime Bouyges", "guid": "67a72cf8-27ec-528a-8f76-1991d9b46f9c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/BKMQ9V/"}, {"code": "T8JC79", "name": "Ghislain Blanchard", "avatar": null, "biography": "Ghislain Blanchard is a research engineer in fluid dynamics at ONERA\u2019s Department of Multi-Physics for Energetics since 2015. His research interests include the development of numerical methods for multi-physics simulations and high performance computing.", "public_name": "Ghislain Blanchard", "guid": "6bd5d606-fc02-5e90-a257-18f4bcce4538", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/T8JC79/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/XSY8DA/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/XSY8DA/", "attachments": []}, {"guid": "9ab0fc08-3dcd-5ccf-bff6-b9cdc106e98e", "code": "7WK8GK", "id": 70865, "logo": null, "date": "2025-10-02T12:20:00+02:00", "start": "12:20", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70865-differentiable-solvers-for-extreme-scale-geophysics-simulations", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/7WK8GK/", "title": "Differentiable solvers for extreme-scale geophysics simulations", "subtitle": "", "track": "Numerical analysis", "type": "Lightning talk", "language": "en", "abstract": "Computational Earth Science relies on modelling to explore complex physical systems that are difficult to observe directly and where data is limited. This is especially crucial in geodynamics and ice sheet dynamics, both of which require large-scale simulations on next-generation HPC systems. These simulations benefit significantly from automated data-driven model constraints, enhancing accuracy and predictive capability.", "description": "We present our recent advancements in large-scale differentiable modelling for high-performance geophysical simulations, focusing on Earth\u2019s largest ice sheets and high-resolution 3D geodynamic models of magmatic systems and mountain formation.\r\n\r\nOur goal is to develop differentiable multi-physics solvers for extreme-scale geophysics simulations, leveraging GPU acceleration and automatic differentiation (AD) with Julia. To achieve this, we integrate Enzyme.jl, an advanced AD tool, enabling highly efficient reverse-mode differentiation at near-theoretical peak performance. This extends accelerated pseudo-transient solvers with differentiable modelling capabilities, significantly enhancing computational efficiency.\r\n\r\nA key aspect of this effort is the enhancement of the [GPU4GEO](https://gpu4geo.org/software/) software stack\u2014a modular ecosystem of Julia-based solvers designed for geophysical simulations. These solvers provide customisable building blocks and optimised subroutines, facilitating rapid development and high-performance execution across multi-core CPUs and AMD/NVIDIA GPUs. By integrating [Enzyme.jl](https://github.com/EnzymeAD/Enzyme.jl), we enable high-performance AD on GPUs, ensuring scalability through MPI-based distributed parallelism on next-generation supercomputers. We will discuss latest advance and report about scalability and simulation results on the [ALPS](https://www.cscs.ch/computers/alps) and [LUMI](https://lumi-supercomputer.eu/about-lumi/) supercomputer at CSCS and CSC, respectively.", "recording_license": "", "do_not_record": false, "persons": [{"code": "ZTATFJ", "name": "Ludovic R\u00e4ss", "avatar": "https://pretalx.com/media/avatars/ZTATFJ_PLThLkL.webp", "biography": "Computational geoscientists with Earth science background. Julia GPU and HPC enthusiast.", "public_name": "Ludovic R\u00e4ss", "guid": "848e7813-344c-5834-aeeb-a0d80906a475", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/ZTATFJ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7WK8GK/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7WK8GK/", "attachments": []}, {"guid": "7b9b921f-5126-5880-b8fa-5a3b8cd6bd9f", "code": "WM8D3B", "id": 70907, "logo": null, "date": "2025-10-02T12:30:00+02:00", "start": "12:30", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70907-scalable-architecture-agnostic-finite-differences-with-chmy-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/WM8D3B/", "title": "Scalable architecture-agnostic finite differences with Chmy.jl", "subtitle": "", "track": "Numerical analysis", "type": "Lightning talk", "language": "en", "abstract": "We present [Chmy.jl](https://github.com/PTsolvers/Chmy.jl), a Julia package for developing scalable finite-difference solvers, with a focus on portability, ease of use, and minimal overhead. Chmy.jl contains modules for working with structured staggered grids, discrete fields, finite-difference and interpolation operators, boundary conditions, and distributed computing. Users can use Chmy.jl in functions that execute both on CPUs and GPUs supported by the JuliaGPU ecosystem.", "description": "Chmy.jl simplifies the implementation of finite-difference solvers for fluid dynamics simulations, with a focus on computational geodynamics and glaciology. It provides tools for writing finite-difference operators in a math-like notation to discretise coupled nonlinear advection-diffusion and Stokes equations. We support various staggered grid layouts, including Arakawa C- and E-grids, and support both Cartesian and curvilinear coordinates.\r\n\r\nDespite leveraging multiple high-level features of Julia, Chmy.jl aims to introduce minimal overhead compared to manually written finite-difference code. All functions typically called in performance-critical paths are non-allocating and inlined. Using Chmy.jl, we scaled a Stokes solver to the full LUMI supercomputer, utilising 21,952 AMD MI250X GPUs and achieving 96% parallel efficiency with an effective resolution of 18,000\u00b3.", "recording_license": "", "do_not_record": false, "persons": [{"code": "FRBW87", "name": "Ivan Utkin", "avatar": "https://pretalx.com/media/avatars/FRBW87_SOwGsLB.webp", "biography": "I'm an applied mathematician working in the field of computational glaciology. My interests include GPU computing, supercomputing, computational fluid dynamics, numerical analysis, to name a few.", "public_name": "Ivan Utkin", "guid": "0c6c77c7-960f-5bcf-af1d-498a9f3b764c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/FRBW87/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/WM8D3B/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/WM8D3B/", "attachments": []}, {"guid": "aefe9df7-5dcc-5999-9b8e-09a70e49955a", "code": "QLLVBE", "id": 70896, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/QLLVBE/benchmark__NO0dx4f.svg", "date": "2025-10-02T12:40:00+02:00", "start": "12:40", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70896-ffts-on-non-equispaced-points-with-nonuniformffts-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/QLLVBE/", "title": "FFTs on non-equispaced points with NonuniformFFTs.jl", "subtitle": "", "track": "Numerical analysis", "type": "Lightning talk", "language": "en", "abstract": "Nonuniform FFT (NUFFT) algorithms enable the efficient evaluation of Fourier series from non-equispaced points. They have a wide range of applications, from medical imaging to the simulation of physical systems. The NonuniformFFTs.jl package provides a highly optimised NUFFT implementation, compatible with CPUs and with various GPU platforms. I will discuss its unique features and the strategies that make it one of the fastest available GPU implementations for large-scale problems.", "description": "The NonuniformFFTs.jl package provides a fast and generic NUFFT implementation in Julia. It is can be parallelised on multiple CPUs using threads, while it can also be run on various GPU platforms thanks to the KernelAbstractions.jl package. It also implements the AbstractNFFTs.jl interface, which allows to easily switch between this and other available NUFFT packages in Julia.\r\n\r\nThis package was initially motivated by the need to compute Fourier series from real-valued data on non-equispaced points. Most existent NUFFT packages only support complex-valued data, leading to unnecessary computations and extra memory usage when data is purely real. This is the case in various applications including molecular dynamics and fluid dynamics simulations. A second motivation was to minimise the cost of successive transforms on different sets of non-uniform points, as these typically change over time in the simulation of physical systems.\r\n\r\nNonuniformFFTs.jl has since evolved into one of the fastest NUFFT implementations available across different languages. On the CPU, it provides comparable performance to the popular FINUFFT library written in C++ (and wrapped in Julia by the FINUFFT.jl package). On GPUs, NonuniformFFTs.jl can currently be considerably (up to 4x) faster than the CUDA-only CuFINUFFT implementation for highly dense three-dimensional problems. This is achieved by taking advantage of GPU shared memory for fast memory accesses, while also avoiding atomic operations on this special kind of memory, which requires to rethink the way the different GPU threads concurrently compute and write to memory.\r\n\r\nIn this talk, I will quickly go through the mathematical definition of the NUFFT and its algorithm. I will then present the main technical difficulties in its practical implementation. Finally, I will discuss the techniques that are used in NonuniformFFTs.jl to accelerate the GPU implementation.", "recording_license": "", "do_not_record": false, "persons": [{"code": "Y7RSST", "name": "Juan Ignacio Polanco", "avatar": "https://pretalx.com/media/avatars/Y7RSST_dHQoHDW.webp", "biography": "I am a research scientist at CNRS and at the [LEGI laboratory](https://www.legi.grenoble-inp.fr/) in Grenoble, France. I work on modelling and simulation of various fluid dynamical systems, with a special interest in turbulent flows and in the use of fast and accurate numerical algorithms. In this context I also develop different tools in Julia. Over the years, I have contributed a small number of open-source Julia packages motivated by my research work. These include [WriteVTK.jl](https://github.com/JuliaVTK/WriteVTK.jl), [PencilArrays.jl](https://github.com/jipolanco/PencilArrays.jl), [PencilFFTs.jl](https://github.com/jipolanco/PencilFFTs.jl), [BSplineKit.jl](https://github.com/jipolanco/BSplineKit.jl), [NonuniformFFTs.jl](https://github.com/jipolanco/NonuniformFFTs.jl) and [VortexPasta.jl](https://github.com/jipolanco/VortexPasta.jl).", "public_name": "Juan Ignacio Polanco", "guid": "103a9a40-4821-5b34-bdd6-857604fa9391", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/Y7RSST/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/QLLVBE/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/QLLVBE/", "attachments": []}, {"guid": "50e1e185-b0ac-5645-8306-c452a41a6392", "code": "GL9P7D", "id": 71194, "logo": null, "date": "2025-10-02T12:50:00+02:00", "start": "12:50", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-71194-divand-a-package-for-smoothing-interpolations-in-n-dimensions-with-topological-features", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/GL9P7D/", "title": "DIVAnd, a package for smoothing interpolations in N dimensions with topological features", "subtitle": "", "track": "Numerical analysis", "type": "Lightning talk", "language": "en", "abstract": "https://github.com/gher-uliege/DIVAnd.jl is a package originally developed for oceanographic data analysis. It allows to calculate gridded and smoothed interpolations from scattered noisy data sets. Over time, features have been added to the tool and the presentation will explain how those can be exploited also in other domains than oceanography and how they compare to other tools available.", "description": "DIVAnd will be explained with some examples from different domains and experience when developing in Julia will be shared, as well the way researchers not used to Julia were able to work with the tools.", "recording_license": "", "do_not_record": false, "persons": [{"code": "HDZ8BT", "name": "Jean-Marie Beckers", "avatar": null, "biography": "Full professor University of Li\u00e8ge\r\nHead of Department Astrophysics, Geophysics and Oceanography\r\nMember of the Royal Academy of Sciences, Belgium", "public_name": "Jean-Marie Beckers", "guid": "d16fcdd5-614c-5b2c-81b7-0fa65d286ef0", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HDZ8BT/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/GL9P7D/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/GL9P7D/", "attachments": []}, {"guid": "38086202-b978-5ecb-a40f-4c2392020600", "code": "MMDZJW", "id": 79779, "logo": null, "date": "2025-10-02T15:30:00+02:00", "start": "15:30", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-79779-dyad-advancing-julia-into-industrial-modeling-and-simulation", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/MMDZJW/", "title": "Dyad: Advancing Julia into Industrial Modeling and Simulation", "subtitle": "", "track": "General", "type": "Sponsor talk", "language": "en", "abstract": "Dyad is a new language for industrial modeling and simulation built by JuliaHub. In this talk we will introduce the audience to Dyad and its capabilities for building high-fidelity hybrid models, integrating acausal continuous simulation with discrete synchronous forms, for optimizing control laws and generating code for embedded devices. We will showcase some of the many features of Dyad, such as generative AI integration with specialized MCP servers for accurate model translation and generation from natural language, scientific machine learning (SciML) features for learning models from data, and integrated tools for control synthesis and analysis. Importantly, we will showcase how the Dyad is built on a foundation of the Julia open source ecosystem, and demonstrate how Julia is a fundamental scripting language for the Dyad experience and how this helps propel the features and usability of the environment beyond traditional modeling tools.", "description": "Sponsor talk by JuliaHub (https://juliahub.com/).", "recording_license": "", "do_not_record": false, "persons": [{"code": "WUWQQ3", "name": "Chris Rackauckas", "avatar": "https://pretalx.com/media/avatars/WUWQQ3_otHw1Wk.webp", "biography": "Dr. Chris Rackauckas is the VP of Modeling and Simulation at JuliaHub, the Director of Scientific Research at Pumas-AI, Co-PI of the Julia Lab at MIT, and the lead developer of the SciML Open Source Software Organization. For his work in mechanistic machine learning, his work is credited for the 15,000x acceleration of NASA Launch Services simulations and recently demonstrated a 60x-570x acceleration over Modelica tools in HVAC simulation, earning Chris the US Air Force Artificial Intelligence Accelerator Scientific Excellence Award. See more at https://chrisrackauckas.com/. He is the lead developer of the Pumas project and received a top presentation award at every ACoP from 2019-2021 for improving methods for uncertainty quantification, automated GPU acceleration of nonlinear mixed effects modeling (NLME), and machine learning assisted construction of NLME models with DeepNLME. For these achievements, Chris received the Emerging Scientist award from ISoP.", "public_name": "Chris Rackauckas", "guid": "6ab02490-d781-5300-94af-5f4ffbefbdff", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/WUWQQ3/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/MMDZJW/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/MMDZJW/", "attachments": []}, {"guid": "f16f7bf4-f3c6-52e9-a127-73ef1a65a41e", "code": "QMYAVP", "id": 70659, "logo": null, "date": "2025-10-02T15:40:00+02:00", "start": "15:40", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70659-dynamical-model-analysis-with-structuralidentifiability-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/QMYAVP/", "title": "Dynamical model analysis with StructuralIdentifiability.jl", "subtitle": "", "track": "SciML", "type": "Lightning talk", "language": "en", "abstract": "When using parametric dynamical models with partial observations, one important task is to infer the parameter values and the states from the measurement data. One of the obstacles to a successful estimation can be the lack of structural identifiability. This means that an unambiguous estimation is impossible due to the very structure of the model, regardless of the quality and quantity of data. We will present StructuralIdentifiability.jl, a Julia package capable of detecting identifiability issues, exploring their nature, and suggesting model transformation improving identifiability.", "description": "Structural identifiability/observability refers to a property of a parametric dynamical model that determines if the model parameters/states can be inferred uniquely in the absence of noise. Lack of structural identifiability may pose problems when calibrating model from data, so assessing this property is an important step in model development. \r\n\r\n[StructuralIdentifiability.jl](https://github.com/SciML/StructuralIdentifiability.jl) is a Julia package developed since 2021 and now allowing different types of identifiability analysis for ODE models with rational dynamics (and also some functionality for discrete-time models). The main functionality includes analyzing different type of identifiability (such as local and global) and, if the model is not structurally identifiable, finding quantities which are nevertheless identifiable and using them to propose a new model. The package has interfaces to interact with ModelingToolkit and Catalyst (the latter by T. Loman).\r\n\r\nThe goal of the presentation will be to present the functionality of the package through a series of case studies.  We will highlight the way different types of analyses can interact with each other and the efficiency/detailedness tradeoffs to consider when dealing with a practical model. We will also explain how to interpret and exploit the results of indentifiability analysis performed by the software. The presentation will conclude with a survey of upcoming features and development directions for the package. We expect that the talk will be of interest for researchers using Julia in the context of dynamical modeling.\r\n\r\nMost of the presented features are based on the first version developed together with R. Dong, C. Goodbrake and H. Harrington and recent works with A. Demin and C. Rackauckas.", "recording_license": "", "do_not_record": false, "persons": [{"code": "G9EREN", "name": "Gleb Pogudin", "avatar": "https://pretalx.com/media/avatars/G9EREN_4chG1I5.webp", "biography": "* PhD in Mathematics from Moscow State University in 2016;\r\n* Postdoc in Johannes Kepler University (Linz, Austria) in 2016-2017 and in New York University in 2017-2019;\r\n* Assistant Professor in Computer Science at the Higher School of Economics (Moscow) in 2019-2020\r\n* Since 2020: Assistant Professor in Computer Science at \u00c9cole Polytechnique, Institute Polytechnique de Paris\r\n\r\n**Research interests:** symbolic computation for differential and difference equations", "public_name": "Gleb Pogudin", "guid": "c535ee89-f630-5471-aaa8-4487a813760c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/G9EREN/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/QMYAVP/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/QMYAVP/", "attachments": []}, {"guid": "966f5664-6d8c-5ff8-a717-b310a45c4512", "code": "3Z8KBZ", "id": 70563, "logo": null, "date": "2025-10-02T15:50:00+02:00", "start": "15:50", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70563-component-based-udes-with-modelingtoolkitneuralnets", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/3Z8KBZ/", "title": "Component based UDEs with ModelingToolkitNeuralNets", "subtitle": "", "track": "SciML", "type": "Lightning talk", "language": "en", "abstract": "Universal Differential Equations (or UDEs for short) have emerged as a novel way to integrate information\r\nfrom experimental data into mechanistic models. In this talk I will present ModelingToolkitNeuralNets, a package intended to help with embedding neural networks inside ModelingToolkit models.", "description": "In this talk we will discuss about how to augment acasual models in the ModelingToolkit framework and about how does the architecture impact the kind of corrections we can make.\r\nThe initial paper on UDEs [1], a methodology is described for systems of differential equations, but in the context of acasual models we start with sets of differential algebraic equations (DAEs) for each component and then a simplification algorithm condenses them in a final system of DAEs that will be solved.\r\n\r\nThis two stage nature of the process naturally divides the possible UDE architectures in two: before structural simplification, where we have we deal with individual components and after simplification where we only have the final simplified system without a direct link to the individual components. We will call these component level UDEs and system level UDEs. In this talk I will mainly focus on component level UDEs.\r\n\r\nThe ModelingToolkitNeuralNets.jl package provides a block component that abstracts a Lux.jl neural network. This block component can be then used as any other component from the ModelingToolkitStandardLibrary. Besides that, we also have a function based interface with the SymbolicNeuralNetwork API.\r\n\r\nIn this talk I will present how to use the package and explore some applications for UDEs in the context of component based models.\r\n\r\n[1] C. Rackauckas et al., \u201cUniversal Differential Equations for Scientific Machine Learning,\u201d arXiv:2001.04385 [cs, math, q-bio, stat], Jan. 2020, Accessed: Feb. 09, 2020. [Online]. Available: http://arxiv.org/abs/2001.04385", "recording_license": "", "do_not_record": false, "persons": [{"code": "DGDNGV", "name": "Fredrik Bagge Carlson", "avatar": "https://pretalx.com/media/avatars/DGDNGV_g61TA1s.webp", "biography": null, "public_name": "Fredrik Bagge Carlson", "guid": "d51dd646-9aed-5eb8-a98c-9c3ef44dbc0c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/DGDNGV/"}, {"code": "HBZFM9", "name": "Sebastian Miclu\u021ba-C\u00e2mpeanu", "avatar": "https://pretalx.com/media/avatars/HBZFM9_fL6acie.webp", "biography": null, "public_name": "Sebastian Miclu\u021ba-C\u00e2mpeanu", "guid": "ffa7b5f6-4c9a-5588-8448-32faf175811d", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HBZFM9/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/3Z8KBZ/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/3Z8KBZ/", "attachments": []}, {"guid": "32aa178a-102a-58f9-85f2-bd9294bdc7f2", "code": "Y9PMHV", "id": 67235, "logo": null, "date": "2025-10-02T16:00:00+02:00", "start": "16:00", "duration": "00:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-67235-optimal-control-in-julia-sciml-s-newest-tooling", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y9PMHV/", "title": "Optimal Control in Julia: SciML's newest tooling", "subtitle": "", "track": "SciML", "type": "Talk", "language": "en", "abstract": "Optimal control is a ubiquitous problem spanning applications of aerospace engineering to chemical processes. Because of that, there are as many softwares for solving optimal control problems as there are applications and domains. Is it possible to unify all optimal control software under one interface? Well that's what SciML does, and we've now entered this domain. In this talk we will demonstrate how ModelingToolkit can be used as a frontend to optimal control that is able to target many of the backend solver methods used throughout various domains. We will also showcase new solvers, built from adaptive BVP solvers, which greatly outperform the standard NLP formulation.", "description": "There are many optimal control softwares that are all built in incompatible ways. The standard SciML trick is to split the problem of developing robust numerical software into two steps: (1) a common interface and (2) new solvers. Using this infrastructure, then it is also easy to benchmark the new solvers to demonstrate the domains in which they outperform the previous methods. Piece of cake, we do that everywhere, and now we're doing that to optimal control.\r\n\r\nIn particular, this general interface needs to be symbolic in order to allow for code generation into the many different solver techniques, some direct methods using ODE solvers, while others codegen to NLP backends like CASADI or JuMP. By making this symbolic interface in ModelingToolkit, we can use the economics of scale afforded by MTK's investment from the systems biology and systems modeling communities in order to achieve a level of development that is generally not available in any open source optimal control tool. We can do things like mix index reduction algorithms with customized codegen and compilers, with specialized CSE algorithms, to get really succinct models that effectively cheat in comparison to commonly used software in this domain.\r\n\r\nBut finally, specialized solvers, built from the BVP solvers, is demonstrated to be a much faster technique than those traditionally done with both direct and indirect methods. The BVP solver uses custom coded linear algebra for fast almost-banded matrices, mixed with a mixed mode automatic differentiation which uses an analytical split between forward and reverse mode, customized GPU acceleration, error control, and adaptivity. We show how this technique greatly outperforms in speed and robustness the basic techniques from CASADI, JuMP, and Pyomo", "recording_license": "", "do_not_record": false, "persons": [{"code": "WUWQQ3", "name": "Chris Rackauckas", "avatar": "https://pretalx.com/media/avatars/WUWQQ3_otHw1Wk.webp", "biography": "Dr. Chris Rackauckas is the VP of Modeling and Simulation at JuliaHub, the Director of Scientific Research at Pumas-AI, Co-PI of the Julia Lab at MIT, and the lead developer of the SciML Open Source Software Organization. For his work in mechanistic machine learning, his work is credited for the 15,000x acceleration of NASA Launch Services simulations and recently demonstrated a 60x-570x acceleration over Modelica tools in HVAC simulation, earning Chris the US Air Force Artificial Intelligence Accelerator Scientific Excellence Award. See more at https://chrisrackauckas.com/. He is the lead developer of the Pumas project and received a top presentation award at every ACoP from 2019-2021 for improving methods for uncertainty quantification, automated GPU acceleration of nonlinear mixed effects modeling (NLME), and machine learning assisted construction of NLME models with DeepNLME. For these achievements, Chris received the Emerging Scientist award from ISoP.", "public_name": "Chris Rackauckas", "guid": "6ab02490-d781-5300-94af-5f4ffbefbdff", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/WUWQQ3/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y9PMHV/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y9PMHV/", "attachments": []}, {"guid": "7b9fb635-a121-567e-82d6-5659d3fb506d", "code": "WLSN8V", "id": 68699, "logo": null, "date": "2025-10-02T17:00:00+02:00", "start": "17:00", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-68699-probing-quantum-properties-of-molecules-with-coulomb-explosion-imaging-and-brute-force-optimization", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/WLSN8V/", "title": "Probing quantum properties of molecules with Coulomb explosion imaging and brute-force optimization", "subtitle": "", "track": "Quantum", "type": "Lightning talk", "language": "en", "abstract": "The principle of Coulomb explosion imaging is simple: use a powerful x-ray pulse to slap a molecule so hard that it (literally) explodes into its constituting atoms. The fragments can then be measured to produce an indirect image of the molecule.\r\n\r\nCurrently, a key question in the Coulomb explosion community is what we can actually learn from these measurements. What information is preserved about the state of the molecule just before it exploded?\r\n\r\nOur simulations show that if the detectors are perfect, we can use this method to image advanced properties of the molecule, namely the high dimensional shape of its ground-state quantum fluctuations. Unfortunately, we can't directly use the experimental data to this end, because it suffers from several key limitations. We circumvent them by fitting a high-dimensional distribution to the data using gradient-based optimization to perform the fit. Then we simply perform our analysis on the fitted model. \r\n\r\nPutting everything together, we demonstrate that the experiment indeed succeeds in imaging subtle quantum properties of the studied molecule.", "description": "Molecules are the building block of pretty much everything. However, this doesn't imply that they are easy to study at the most fundamental level. In fact, even when they are isolated, they remain complex quantum systems, challenging to model theoretically and to image experimentally.\r\n\r\nTo address this challenge, new scientific facilities (called XFEL for X-ray Free Electron Laser) were built to provide scientists with absurdly powerful pulses of x-ray. One of the uses of the high power delivered at these instruments is to induce clean explosions of molecules. By destroying the molecule and measuring the produced fragments, one can infer some information about the molecule just-before-explosion, a technique known as Coulomb explosion imaging. The method is currently gaining popularity, on the one hand because the infrastructures are rapidly improving, and on the other hand because it is one of the few contenders that may one day be able to produce images and movies of arbitrary molecules, a feat currently out of reach for any method.\r\n\r\nA key problem is to generalize Coulomb explosion imaging to larger molecules. In general, not all fragments can be detected simultaneously, leading to measurements with a lot of missing information, which complicate the interpretation of the data.\r\n\r\nWe have developed a method to reconstruct this information, by fitting a high-dimensional Gaussian distribution to the data, through a gradient-descent optimization of the fitted model. We defined the loss function of the problem in such a way that it is not affected by the information missing from the measurements, making it suitable to analyze the experimental data.\r\n\r\nThen, with the help of simulation of the Coulomb explosion of a molecule, we were able to link the measurements to fundamental properties of the molecule, namely the shape of its quantum ground-state fluctuations.\r\n\r\nA large part of the talk will be dedicated to introducing the concept of Coulomb explosion itself, and to detail the challenge it faces in terms of data analysis. Then I will describe the reconstruction method that I devised, with an emphasis on how the specifics of Julia were helping me. In particular, I'll show the impact that package like Tullio.jl, Zygote.jl and Optim.jl had on the project. Finally, I'll briefly conclude on the physical relevance of my work in relation to the imaging of quantum property of a large molecule.", "recording_license": "", "do_not_record": false, "persons": [{"code": "KMMWMD", "name": "Beno\u00eet Richard", "avatar": "https://pretalx.com/media/avatars/KMMWMD_wAt8C88.webp", "biography": "Doctor in theoretical physics, working mostly in simulation and data analysis with experimentalists blowing up tiny things in large facilities.\r\n\r\nWorked on IntervalArithmetic.jl for my master thesis and stuck around.\r\n\r\nMade the questionable life choice of writing a LaTeX engine in julia.\r\n\r\nFirst julia version used: v0.4", "public_name": "Beno\u00eet Richard", "guid": "2eb1a124-3630-547c-bb4d-212df7fed890", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/KMMWMD/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/WLSN8V/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/WLSN8V/", "attachments": []}, {"guid": "238d0f3f-a937-5ed8-91f0-881c29556be5", "code": "XTZEPJ", "id": 72154, "logo": null, "date": "2025-10-02T17:10:00+02:00", "start": "17:10", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-72154-quantum-control-with-julia-optimal-pulses-and-lie-algebra-methods", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/XTZEPJ/", "title": "Quantum Control with Julia: Optimal Pulses and Lie Algebra Methods", "subtitle": "", "track": "Quantum", "type": "Lightning talk", "language": "en", "abstract": "Quantum technologies exploit small-scale systems to advance computation, communication, and metrology. In this work we develop theory and algorithms for optimal pulse control using indirect methods based on Pontryagin maximum principle. Simulations and numerical solution are obtained by means of the Julia package [OptimalControl.jl](https://control-toolbox.org/OptimalControl.jl) to couple direct and indirect methods. We also use Julia to implement efficient algorithms to generate the free Lie algebra in the Hall basis to assess controllability of quantum systems.", "description": "## Quantum control with Julia: Optimal pulses and Lie algebra methods\r\n\r\n### D. Tinoco[^1], I. Beschastnyi[^1], J.-B. Caillau[^2]\r\n\r\n### Abstract\r\n\r\nQuantum technologies exploit small-scale systems to advance computation, communication, and metrology. In this work we develop theory and algorithms for optimal pulse control using indirect methods based on Pontryagin maximum principle. Simulations and numerical solution are obtained by means of the Julia package [OptimalControl.jl](https://control-toolbox.org/OptimalControl.jl) to couple direct and indirect methods. We also use Julia to implement efficient algorithms to generate the free Lie algebra in the Hall basis to assess controllability of quantum systems.\r\n\r\n### Description\r\n\r\nQuantum technologies exploit small-scale physical systems and are expected to revolutionize computing, communication, and metrology fields. However, these systems are highly sensitive and short-lived, making them difficult to control. Our goal is to develop a theoretical and computational framework for efficiently steering quantum systems using pulse controls. This type of control is currently one of the fastest and most practical way to manipulate real-world quantum systems. Our approach is grounded in Pontryagin\u2019s maximum principle, which helps us characterize time-optimal controls. To ensure that we can actually steer a system between desired states, we analyse controllability. This is done by generating the free Lie algebra of the controlled system, using a Hall basis built from Lyndon words. We implement our methods in Julia, both for the controllability analysis and for the numerical resolution with direct transcription and indirect methods (*aka* shooting) thanks to the package [OptimalControl.jl](https://control-toolbox.org/OptimalControl.jl) [2]. As a first use case, we successfully reproduce known optimal pulse sequences from the literature [1]. We also verify the controllability of quantum systems generating the associated free Lie algebra. In particular, we leverage some tools from the package [PauliStrings.jl](https://paulistrings.org) [3] that performs fast group operations by efficiently encoding Pauli strings. \r\n\r\n1. Ansel, Q.; Dionis, E.; Arrouas, F.; Peaudecerf, B.; Gu\u00e9rin, S.; Sugny, D. Introduction to theoretical and experimental aspects of quantum optimal control. *J. Phys. B* **57** (2024), no. 13, 133001.\r\n2. Caillau, J.-B., Cots, O., Gergaud, J., Martinon, P., & Sed, S. OptimalControl.jl: a Julia package to model and solve optimal control problems with ODE's. [doi.org/10.5281/zenodo.13336563](https://doi.org/10.5281/zenodo.13336563)\r\n3. Loizeau, N.; Peacock, C. J.; Sels, D. Quantum many-body simulations with PauliStrings.jl. *SciPost Phys. Codebases* **54** (2025)\r\n\r\n[^1]: Universit\u00e9 C\u00f4te d'Azur, Inria, CNRS, LJAD\r\n\r\n[^2]: Universit\u00e9 C\u00f4te d'Azur, CNRS, Inria, LJAD", "recording_license": "", "do_not_record": false, "persons": [{"code": "NHSCFX", "name": "David Tinoco", "avatar": null, "biography": "I am currently a PhD student at Inria, supervised by Jean-Baptiste Caillau and Ivan Beschastnyi. My research topic is in quantum control theory, with a focus on nearly optimal pulse control of quantum systems. The aim of this project is to develop both theoretical and numerical methods for the fast and efficient manipulation of quantum systems, which are known to be highly sensitive to noise and decoherence.", "public_name": "David Tinoco", "guid": "3879bde2-c84f-5303-9aa0-0d5b9dfdb4c6", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/NHSCFX/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/XTZEPJ/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/XTZEPJ/", "attachments": []}, {"guid": "53d51e24-37e3-57f4-8171-a6d9ecc2a6b5", "code": "Q73CNG", "id": 70599, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/Q73CNG/rimu_U0tMEzd.png", "date": "2025-10-02T17:20:00+02:00", "start": "17:20", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70599-sampling-the-eigenstates-of-an-infinite-dimensional-matrix-with-rimu-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/Q73CNG/", "title": "Sampling the eigenstates of an infinite dimensional matrix with Rimu.jl", "subtitle": "", "track": "Quantum", "type": "Lightning talk", "language": "en", "abstract": "Quantum many-body problems famously scale very badly and sometimes it cannot be avoided to work with a Hilbert space of infinite dimension. Rimu.jl can represent a Hamiltonian as a sparse matrix and implements a projector Monte Carlo method that can sample the ground and low-lying eigenstates, even when the matrix size is huge. This talk will introduce the Rimu.jl package and focus on updates since the JuliaCon24 talk. The highlights will include a new link to the quantum chemistry package ElemCo.jl and a bias-free taming of the Monte Carlo sign problem in the Fr\u00f6hlich polaron model with a formally and practically infinite Hilbert space.", "description": "Rimu.jl aims to be an extensible platform for quantum many-body calculations. The problem of finding the energy and properties of the ground and low-lying excited quantum states is central to many important questions in quantum chemistry and condensed-matter physics. Formally these problems map into the linear algebra problem of finding the eigenvectors and eigenvalues of a large matrix. The size of the matrix scales very badly, typically exponentially, with the size of the physical system. Rimu.jl implements matrix-based and matrix-free exact diagonalisation, which has seen a significant upgrade and improved speed since last year's JuliaCon24 talk. When the matrix dimension is so large that eigenvectors can no longer be stored in memory, Rimu.jl can still help and sample the eigenstate with projector Monte Carlo, which is implemented in the flavour of full configuration interaction quantum Monte Carlo (FCIQMC). Rimu.jl is written in pure Julia and highly extensible. Computations are fully parallelised using native threading and MPI on HPC clusters.\r\n\r\nIn this talk I will highlight a recent application where the Monte Carlo sign problem for calculating excited states could be overcome by walker annihilation in the Fr\u00f6hlich polaron model with a formally infinite linear dimension. I will further report on recent progress with linking Rimu.jl with the quantum chemistry package ElemCo.jl and the JuliaMolSim ecosystem and implementing a molecular Hamiltonian.", "recording_license": "", "do_not_record": false, "persons": [{"code": "YGNY3T", "name": "Joachim Brand", "avatar": "https://pretalx.com/media/avatars/YGNY3T_dDmXhhV.webp", "biography": "Joachim is a theoretical and computational physicist interested in quantum many-body physics, quantum simulation, and quantum technology.", "public_name": "Joachim Brand", "guid": "832002f5-971f-5413-a6f2-5440e09560f2", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/YGNY3T/"}], "links": [{"title": "Rimu.jl github package repository", "url": "https://github.com/RimuQMC/Rimu.jl", "type": "related"}], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Q73CNG/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Q73CNG/", "attachments": []}, {"guid": "55f5ac61-bc95-5bd4-837b-c28e5a0cdf23", "code": "PCHDZN", "id": 72201, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/PCHDZN/Screenshot_uyXSWbU.jpg", "date": "2025-10-02T17:30:00+02:00", "start": "17:30", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-72201-gpu-accelerated-tractography-in-julia", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/PCHDZN/", "title": "GPU-Accelerated Tractography in Julia", "subtitle": "", "track": "Neuroscience", "type": "Lightning talk", "language": "en", "abstract": "Tractography.jl is a Julia package for reconstructing numerical representations of white matter (WM) fiber pathways in the brain from magnetic resonance imaging (MRI) data. It supports multiple tractography algorithms and includes a visualization backend based on Makie.jl for exploring these complex datasets interactively.", "description": "We present new algorithms for reconstructing white matter (WM) fiber pathways in the brain from MRI data, implemented entirely in Julia. Unlike conventional methods, our approach is optimized for GPU computation, achieving significant speedups while maintaining flexibility. The pipeline integrates a custom visualization backend built on Makie.jl, enabling interactive exploration of complex tractography data.\r\n\r\nIn this talk, I will introduce the theoretical background of tractography, including spherical harmonics-based modeling, describe the core algorithms, and discuss implementation details. I will also outline current limitations of the GPU implementations in both tracking and visualization, and suggest directions for future improvements.", "recording_license": "", "do_not_record": false, "persons": [{"code": "PGWM93", "name": "Romain VELTZ", "avatar": null, "biography": null, "public_name": "Romain VELTZ", "guid": "5c3c82c4-fe04-5808-a9c7-3eb1d2476334", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PGWM93/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/PCHDZN/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/PCHDZN/", "attachments": []}, {"guid": "ee77e943-b186-54f6-a99b-c7b657804997", "code": "NDHNX9", "id": 70908, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/NDHNX9/logo_Unfol_qC5jIp4.png", "date": "2025-10-02T17:40:00+02:00", "start": "17:40", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70908-the-unfold-family-ecosystem-for-rerp-eeg-analyses", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/NDHNX9/", "title": "The Unfold Family: Ecosystem for rERP-EEG analyses", "subtitle": "", "track": "Neuroscience", "type": "Lightning talk", "language": "en", "abstract": "The Unfold family is a set of time-series analysis, visualization, and simulation toolboxes built around the regression ERP (rERP) framework for EEG analysis. That is, the application of multiple regression models to time series data. While all packages have a focus on EEG data, they can typically be easily applied to pupil dilation, skin conductance, or even fMRI data.", "description": "The Unfold family consists of seven packages: `Unfold.jl`, `UnfoldMixedModels.jl`, `UnfoldMakie.jl`, `UnfoldSim.jl`, `UnfoldBIDS.jl`, `UnfoldDecode.jl` and `UnfoldStats.jl`, together offering a very broad set of EEG analysis use cases. \r\n\r\nWe can use these tools to analyze complex experimental designs, especially in naturalistic settings. There we often struggle with (non-)linear confounds and temporal overlapping brain responses (see [Dimigen & Ehinger 2021](https://doi.org/10.1167/jov.21.1.3)). Analyzing such data proofs challenges on all levels, which we try to address with dedicated packages: The main analysis packages are `Unfold.jl` for rERP, and `UnfoldMixedModels.jl` to extend the analyses to linear mixed models. Method-validity can be tested using simulations (`UnfoldSim.jl`), complex data can be explored and visualized (`UnfoldMakie.jl`), insights can be gained via statistics (`UnfoldStats.jl`). To make all steps easily reproducible, we offer `UnfoldBids.jl`, a pipeline-solution based on the \"Brain Imaging Data Structure\" standard. Finally, `UnfoldDecode.jl` offers a set of experimental procedures to apply more machine-learning approaches to such data, decoding brain states or conditions.\r\n\r\nAll packages are tested and well documented, and offer plenty of tutorials for beginners accessible via the [`UnfoldDocs` website](https://unfoldtoolbox.github.io/UnfoldDocs/main/).\r\n\r\n\r\nIn the talk I will present some highlights of my research in that field over the last 10 years. I'll present some exemplary results we performed using our Unfold-Family. And, if time permits, I can share some lessons learned when developing in a diverse field of non-computerscientists, and translating a toolbox from matlab to Julia.", "recording_license": "", "do_not_record": false, "persons": [{"code": "GMJFMU", "name": "Benedikt Ehinger", "avatar": "https://pretalx.com/media/avatars/GMJFMU_JCZiQuT.webp", "biography": "Assistant Professor for **Computational Cognitive Science**\r\nUniversity of Stuttgart", "public_name": "Benedikt Ehinger", "guid": "0cf214d7-1b13-54f8-b29d-899ac12122c6", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/GMJFMU/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NDHNX9/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NDHNX9/", "attachments": []}, {"guid": "ec885b86-4207-5ce9-9231-0e0f58309a44", "code": "VVYMWH", "id": 70918, "logo": null, "date": "2025-10-02T17:50:00+02:00", "start": "17:50", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70918-unfoldmakie-jl-advanced-visualization-of-regression-event-related-potentials", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/VVYMWH/", "title": "UnfoldMakie.jl: Advanced Visualization of (regression) Event Related Potentials", "subtitle": "", "track": "Neuroscience", "type": "Lightning talk", "language": "en", "abstract": "In this 10-minute talk, we will explore how UnfoldMakie.jl visualizes electrical activity of the brain recorded from the scalp. It works with electroencephalographic (EEG) data and such techniques as event-related potentials (ERPs) and regression-based ERPs (rERPs). UnfoldMakie integrates ERP and rERP analysis from Unfold.jl with interactive, customizable plotting from Makie.jl. This session will demonstrate how to create interactive and insightful visualizations of scalp electrical activity in Julia.", "description": "UnfoldMakie.jl is a Julia package designed to simplify the visualization of EEG data, ERP and regression ERP (rERP), focusing on flexibility, speed, and interactivity (Mikheev & Ehinger, 2025). UnfoldMakie combines the powerful techniques for ERP and rERP analysis  from Unfold.jl with the high-performance plotting capabilities of Makie.jl and the flexible, declarative plotting approach of AlgebraOfGraphics.jl. This integration provides an intuitive interface for creating detailed, customizable plots that help researchers interpret and present their data more effectively. \r\nThis work builds on the research presented in Mikheev et al. (2024), where we conducted a user study to analyze the usage of existing EEG tools, gathering feedback on user preferences, challenges, and needs. The insights from this study informed the design of UnfoldMakie.jl, ensuring it addresses real-world user demands and overcomes common difficulties faced with other EEG software tools.\r\nIn this session, we'll walk through the core features and functionality of UnfoldMakie, including:\r\n    \u2022 Getting Started: Setting up UnfoldMakie and importing data. Glossary of ERP plot types and how they were implemented in the package. Comparison with other non-Julia tools.\r\n    \u2022 Visualizing Data: Creating interactive ERP and EEG plots for the most of existing ERP plot types, from simple time-series graphs to complex topographic plots.\r\n    \u2022 Regression ERP (rERP) Analysis: Introducing rERP and demonstrating how UnfoldMakie can handle regression-based ERP analysis, making it easier to visualize the effects of predictors on event-related potentials.\r\n    \u2022 Customization: Customizing plots to fit specific research needs, with options for multi-plot layouts, dynamic data adjustments, and visual tweaks.\r\n    \u2022 Future Directions: Discussing the roadmap for UnfoldMakie, including planned features for more flexible visualization options, and the addition of various uncertainty visualization techniques.\r\nThis talk will cater to those working in Julia with time-series data in fields such as neuroscience or cognitive sciences. \r\n\r\nCitations:\r\n    \u2022 Mikheev, V., Skukies, R., & Ehinger, B. V. (2024). The Art of Brainwaves: A survey on event-related potential visualization practices. Aperture Neuro, 4. https://doi.org/10.52294/001c.116386\r\n    \u2022 Mikheev, V., & Ehinger, B. V. (2025). UnfoldMakie.jl: EEG/ERP visualization package. Journal of Open Source Software, 10(105), 7560. https://doi.org/10.21105/joss.07560", "recording_license": "", "do_not_record": false, "persons": [{"code": "8AWCAG", "name": "Vladimir Mikheev", "avatar": "https://pretalx.com/media/avatars/8AWCAG_mLNTnXr.webp", "biography": "I work at University of Stuttgart and do visualizations for neuroscience.", "public_name": "Vladimir Mikheev", "guid": "74712cd8-d761-5792-82bc-f88d893d1b22", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/8AWCAG/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VVYMWH/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VVYMWH/", "attachments": []}], "Coffee room": [{"guid": "0ffbf124-0ed2-57bf-9756-da7f6f497330", "code": "F7BTWX", "id": 70668, "logo": null, "date": "2025-10-02T11:00:00+02:00", "start": "11:00", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70668-monte-carlo-simulations-with-paralleltemperingmontecarlo-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/F7BTWX/", "title": "Monte Carlo simulations with ParallelTemperingMonteCarlo.jl", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Our package implements the parallel-tempering Monte Carlo method to study the thermodynamics of physical systems. Several copies of the system at different temperatures, randomly initialized, are evolved in parallel and allowed to exchange information of their current state. We make extensive use of Julia\u2019s dynamic type system and multiple dispatch to unify implementations for different ensembles, interaction potentials and boundary conditions; multi-threading is used for parallelization.", "description": "We are studying solid-solid and solid-liquid phase transitions in atomic/molecular systems by means of the parallel tempering Monte Carlo method - also known as replica-exchange Markov chain Monte Carlo. In order to obtain the needed thermodynamic information several Markov chains of atomic/molecular configurations are generated simultaneously over the temperature range of interest. New configurations are generated by particle displacement, volume change or atomic exchange moves depending on the ensemble and system of choice. To prevent simulations at low temperatures to become stuck in local minima of the potential energy landscape, configurational exchanges between neighboring temperatures are attempted occasionally. New interesting regions in configurational space found at high temperatures where trajectories can move freely across potential barriers can thus be communicated \u201cdown\u201d to lower temperatures. Energy histogram information is stored during the simulations and is checked for sufficient histogram overlap between neighboring temperatures to ensure the efficiency of the exchange mechanism.\r\n\r\nUsing Julia\u2019s rich type system, we implemented different types of canonical and isobaric-isothermal ensembles with corresponding move strategies. Interaction potentials from extended Lennard Jones potentials, look-up tables to embedded-atom models are already implemented and an interface to a Fortran implementation of an atomic neural network potential exists. Besides the histogram information, radial distribution functions, sampled configurations and exchange statistics are stored.\r\n\r\nWe are also developing a Julia tool box for post-processing of the simulation results. At present we are using a multi-histogram analysis for calculation of the partition function allowing access to accurate internal energy, entropy and heat capacity information. Further, we are implementing structural analysis methods to get insight into how efficient configurational space is explored. For this, we perform energy minimization of atomic configurations saved during our simulations followed by a common-neighbor analysis and cluster similarity measures. The similarity scores allow for a compact visualization of the configurational space sampled and can be compared with a corresponding visualization of the potential energy surface gained by the basin-hopping algorithm.\r\n\r\nAfter a short introduction of the Monte Carlo method we will present our implementation strategy while showcasing melting simulations for small neon clusters in strong magnetic fields and bulk argon under extremely high pressures.", "recording_license": "", "do_not_record": false, "persons": [{"code": "PWUFUZ", "name": "Elke Pahl", "avatar": "https://pretalx.com/media/avatars/PWUFUZ_Q046eot.webp", "biography": "I originally studied chemistry in Heidelberg, Germany gaining a PhD in Theoretical Chemistry. My extended postdoc time was spent in the US, back in Germany (Dresden) and then at Massey University Albany, transitioning more and more into physics while also taking care of and time off for the education of my children. I am now a Senior Lecturer in physics at the University of Auckland. My research lies in computational modelling from more fundamental topics like melting of matter in extreme conditions or exploring ground and excited states of quantum phases to more applied modelling of percolating carbon-elastomer composites.", "public_name": "Elke Pahl", "guid": "331f9abd-8c31-5160-bcb5-8efbb6f407c7", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PWUFUZ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/F7BTWX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/F7BTWX/", "attachments": []}, {"guid": "ef2221d0-3f60-5750-9a30-83ccd14dde4b", "code": "UDDRRM", "id": 70915, "logo": null, "date": "2025-10-02T11:03:00+02:00", "start": "11:03", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70915-component-based-hybrid-model-simulations-beyond-modelingtoolkit", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/UDDRRM/", "title": "Component-based hybrid model simulations: Beyond ModelingToolkit", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "ModelingToolkit.jl provides many benefits as a symbolic component-based modelling language, including efficient code generation and structural simplifications. However, it is limited in its extensibility outside of the SciML ecosystem. Being able to integrate and connect ModelingToolkit systems with other modelling tools in Julia like Agents.jl and with external libraries would open up a range of hybrid modelling options to the Julia community.", "description": "ModelingToolkit.jl is an acausal component-based modelling language which provides many benefits including efficient code generation, structural simplifications and a simple system for connecting models. However, it is limited in its extensibility outside of the SciML ecosystem. Modelling tools like Agents.jl and calls to external libraries can be challenging to define and limited in functionality, if possible at all. We introduce a component-based simulation tool, to complement ModelingToolkit\u2019s component-based modelling language, which allows finer control for coupled simulations, allowing arbitrary Julia simulations to be connected to the ModelingToolkit ecosystem. \r\nWe achieve this through a general component and integrator interface, allowing ModelingToolkit generated systems to be simulated, coupled together with other components and integrators, such as those predefined for Agents.jl, or defined by the user through implementing a simple integrator interface. We can utilise the graph of these connections between components to improve the simulation efficiency, such as computing components in parallel, or replacing components with surrogates.", "recording_license": "", "do_not_record": false, "persons": [{"code": "7QEJDL", "name": "Matt Owen", "avatar": "https://pretalx.com/media/avatars/7QEJDL_gTpiBJY.webp", "biography": "I am a mathematical biologist from the University of Bristol, UK. I have previously worked on efficient parameterisation of cardiac ion channel models, uncertainty quantification and models of blood coagulation. I currently work on the software pillar of the EEBio synthetic biology grant.", "public_name": "Matt Owen", "guid": "0a1f2fc0-774e-53f5-9071-c9e84368f17c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/7QEJDL/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UDDRRM/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UDDRRM/", "attachments": []}, {"guid": "6fc18d2a-f8a1-55c2-9a60-a44702b10459", "code": "BXBAEY", "id": 72092, "logo": null, "date": "2025-10-02T11:06:00+02:00", "start": "11:06", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72092-clustering-for-optimization-linear-program-reductions-with-tulipaclustering-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/BXBAEY/", "title": "Clustering for Optimization: Linear Program Reductions with TulipaClustering.jl", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "*TulipaClustering.jl* introduces a greedy hull clustering method designed to better preserve constraint-binding extremes in optimization problems, especially within energy systems modeling. This approach yields more accurate reduced models for improved downstream optimization performance.", "description": "Clustering is a popular tool for reducing the size of datasets before solving large-scale optimization problems\u2014particularly in energy systems modeling. Common methods like *k*-means and *k*-medoids select points near the center of clusters, aiming to represent average behavior in the data. These centroids are then used as representative data points to build the reduced model. However, this approach often misses the points that matter most: the extremes that are more likely to be constraint-binding in the full optimization problem.\r\n \r\n*TulipaClustering.jl*, takes a different approach: a (greedy) hull clustering algorithm tailored to reductions of optimization problems. Instead of maximizing intra-cluster similarity, we prioritize covering the convex hull of the input data with a small set of representative points. These representatives are more likely to approximate the binding regions of the original problem\u2019s feasible space, leading to tighter approximations and better decisions downstream. Additionally, we use blended weights to approximate interior points as mixtures of the representatives. This allows the reduction to preserve key structural features of the original problem, while significantly reducing dimensionality.\r\n \r\nIn this talk, we describe the motivation, algorithm, and implementation behind *TulipaClustering.jl*, and show how it integrates into energy system models. We share practical results from energy-system optimization case studies, where greedy hull clustering outperforms conventional clustering methods in downstream optimization quality without sacrificing runtime. For energy modelers and optimization practitioners, *TulipaClustering.jl* bridges the gap between statistical model reduction and optimization-ready problem formulation.", "recording_license": "", "do_not_record": false, "persons": [{"code": "MQNYYE", "name": "Diego Tejada", "avatar": null, "biography": null, "public_name": "Diego Tejada", "guid": "c6d00cf5-9c18-566f-9eb6-019b3d55c170", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/MQNYYE/"}, {"code": "XRHXWS", "name": "Grigory Neustroev", "avatar": null, "biography": "Grigory Neustroev as a postdoctoral researcher at Delft University of Technology. His research interests lie at the intersection of AI and optimization, with application to energy problems.", "public_name": "Grigory Neustroev", "guid": "3c3ebbdd-4b5f-58e7-876c-a0cea594eda5", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/XRHXWS/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/BXBAEY/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/BXBAEY/", "attachments": []}, {"guid": "d6640d0b-0acb-5cb7-b86f-e98d5361ca12", "code": "NM83XC", "id": 70972, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/NM83XC/Capture_de_ZRYwsgC.png", "date": "2025-10-02T11:09:00+02:00", "start": "11:09", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70972-climflows-toward-an-ecosystem-of-composable-packages-for-the-simulation-of-climate-relevant-flows", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/NM83XC/", "title": "ClimFlows: toward an ecosystem of composable packages for the simulation of climate-relevant flows", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "ClimFlows is an ongoing effort to develop an ecosystem of composable and extensible packages from which one can compose a variety of climate-relevant numerical models, especially for oceanic and atmospheric circulation, from toy models to realistic solvers. \r\nWhile implementation is important, interfaces (APIs) are essential for the composability and extensibility of such an ecosystem. The talk will exemplify how building blocks of the ecosystem (spectral harmonics transforms, time integration, thermodynamics...) can be composed into two global atmospheric solvers, focus on a sample of packages and their interfaces, and discuss future work and interaction with the Julia community.", "description": "Established or more recent efforts to develop numerical models of the atmosphere or ocean are to a large extent monolithic, in the sense that a single package offers a complete but tightly tangled set of features, in an all-or-nothing manner.\r\n\r\nClimFlows is an ongoing effort that bets on a different approach: develop an ecosystem of composable and extensible packages:\r\n- from which one can compose a variety of climate-relevant numerical models, especially for oceanic and atmospheric circulation, from toy models to realistic solvers\r\n- from which one can cherry-pick specific functionality without having to buy into the whole ClimFlows ecosystem\r\n- whose performance is competitive on laptops as well as HPC clusters, on CPUs and GPUs\r\n - whose composability builds upon state-of-the-art understanding of the mathematical and physical structure of geophysical models\r\n\r\nThe talk will exemplify how building blocks of the ecosystem can be composed into two global atmospheric solvers - finite-difference and spectral - with flexible thermodynamics. \r\n\r\nWe will emphasize a few packages of the ecosystem that are more mature or can be of interest for the wider community: our thermodynamics package, a lightweight performance portability layer, and a package for spectral harmonics transforms with custom adjoints for Zygote and Enzyme. While we comment to some extent on implementation, our main focus will be on the interfaces (APIs), which are the keys to a composable and extensible ecosystem.\r\n\r\nTo conclude, we will sketch a roadmap, how our effort could benefit the community, and potential areas for collaboration with computational scientists, applied mathematicians, and AI experts.", "recording_license": "", "do_not_record": false, "persons": [{"code": "8REG8A", "name": "Thomas Dubos", "avatar": null, "biography": null, "public_name": "Thomas Dubos", "guid": "3e9773e9-024a-54e9-bd20-2f669b19f76f", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/8REG8A/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NM83XC/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NM83XC/", "attachments": []}, {"guid": "acf6cf9e-dcd7-593a-8aa4-882129d33875", "code": "DRAK8T", "id": 70644, "logo": null, "date": "2025-10-02T11:12:00+02:00", "start": "11:12", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70644-circuit-model-discovery-algorithm-for-supercapacitors", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/DRAK8T/", "title": "Circuit Model Discovery Algorithm for Supercapacitors", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "A supercapacitor is a complex charge storing device which relies on multiple physical phenomena, namely electrical double layer formation and pseudocapacitance, that are to this day not so well understood. When theory stagnates, so does innovation. Where trial and error give results, it could give faster results if the trial could be guided by knowledge. Part of the process in trying to understand how a supercapacitor works is finding an equivalent circuit model, acting as a sort of a rigorous guide through uncharted territory using charted territory in the form of already known circuit elements which describe, to some extent, the physical phenomena behind a supercapacitor. Therefore, this is what the presented work is concerned with, using Julia to find a circuit model from acquired supercapacitor data.", "description": "Current equivalent circuit model discovery algorithms require the user to input a plausible equivalent circuit based on the measured data they have, which is Bode and Nyquist plots from electrochemical impedance spectroscopy (EIS) measurements, and the software begins an optimization process of fitting the data with the theoretical model corresponding to the equivalent circuit candidate until there is a close enough match between them. This method leaves a lot to be desired, for if this particular circuit could fit the data, so could other a million mathematically equivalent circuits, therefore making certainty over the result quite impossible. What the algorithm developed by me plans to do is gather as much information as possible from the data itself first, automatically, and through that information have the said algorithm suggest some possible circuit configurations that fit the physical equations. Therefore, a main stage of parameter discovery and one of parameter discrimination would be introduced for a better accuracy for the results. Afterwards, an optimization stage will use the discovered parameters and the theoretical model for the circuit configuration in order to fit it to the existing data. Having some initial values at the start of optimization will guarantee the model will be more likely to converge to the actual circuit values that would accurately model the respective supercapacitor measured. This work proposes on improving on the circuit model discovery package I have previously developed which currently models part of the supercapacitor\u2019s behaviour using Optimization.jl (https://github.com/Ir1n-a/NyquistCircuitModeling.jl).", "recording_license": "", "do_not_record": false, "persons": [{"code": "RBRPFZ", "name": "Irina-Nicoleta Bratosin", "avatar": "https://pretalx.com/media/avatars/RBRPFZ_1EtGDVr.webp", "biography": "I am a PhD student in physics, working as a researcher with my main specialty being supercapacitors and electric measurements.  I got into Julia because I believe it's the right programming language for the scope of my research, and, well, I also like it more than other programming languages, so there's that.", "public_name": "Irina-Nicoleta Bratosin", "guid": "a5a85066-d5e3-528e-9eaf-06d732e1b293", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/RBRPFZ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/DRAK8T/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/DRAK8T/", "attachments": []}, {"guid": "7784cd90-8902-57fd-97f2-78017b59c2ea", "code": "LAXF3S", "id": 72339, "logo": null, "date": "2025-10-02T11:15:00+02:00", "start": "11:15", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72339-accurate-ground-state-computation-of-atoms-and-ions-with-extended-kohn-sham-models", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/LAXF3S/", "title": "Accurate Ground-State Computation of Atoms and Ions with Extended Kohn\u2013Sham Models", "subtitle": "", "track": "Posters", "type": "Poster", "language": "fr", "abstract": "We present a Julia implementation of extended Kohn-Sham models (LDA/LSDA) for computing ground states of atoms and ions using high-precision finite element methods. Originally developed to investigate the existence of negative ions, the code also serves as a versatile tool for testing new density functionals, studying numerical precision issues.", "description": "This package aims to compute the ground state of atoms and ions using extended Kohn\u2013Sham models (LDA/LSDA). These models require solving a nonlinear eigenvalue PDE in 3D, where the nonlinearity comes from a potential depending on a finite number of eigenfunctions. As is standard in computational chemistry, a fixed-point iterative scheme is used, alternating with the resolution of the eigenvalue problem.\r\n\r\nThanks to spherical symmetry and the use of spherical harmonics, the eigen value problem can be reduced to a family of radial equations in 1D. The package implements a high-precision finite element method for solving these equations, using polynomials of degree up to 20 on geometric meshes. This allows for highly accurate results with fewer mesh points, making it more efficient than low-order approaches with refined meshes.\r\n\r\nThe code supports both double and quadruple floating-point precision. Such high precision is essential to resolve subtle numerical questions left open in the literature. For example, in certain extended Kohn\u2013Sham models, it remains unclear whether the outermost orbital energies for some atoms are exactly zero or merely very small but negative \u2014 a distinction that can now be investigated numerically.\r\n\r\nThe package includes support for a wide range of exchange-correlation functionals and offers two iterative solvers: ODA and a quadratic method. \r\n\r\nLink of the code : https://github.com/Theozeud/AtomicKohnSham", "recording_license": "", "do_not_record": false, "persons": [{"code": "7RN3AF", "name": "Duez Th\u00e9o", "avatar": "https://pretalx.com/media/avatars/7RN3AF_NReYmgv.webp", "biography": "Doctorant en premi\u00e8re ann\u00e9e sous la direction d'\u00c9ric Canc\u00e8s et Mathieu Lewin", "public_name": "Duez Th\u00e9o", "guid": "5d75d7c5-9af2-5e2d-8e9f-10c41a107085", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/7RN3AF/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LAXF3S/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LAXF3S/", "attachments": []}, {"guid": "f908438d-d714-555c-a2a6-db6f61cfd53f", "code": "YFYL3K", "id": 70921, "logo": null, "date": "2025-10-02T11:18:00+02:00", "start": "11:18", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70921-a-fourier-continuation-fc-framework-for-high-order-pde-solvers", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/YFYL3K/", "title": "A Fourier continuation (FC) framework for high-order PDE solvers", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "The use of Fourier series for high-order trigonometric interpolation can be found in a plethora of applications in science and engineering. Here we present a framework for constructing and testing recently-introduced extension operators that can enable such representations for non-periodic functions (while avoiding the ringing effect).  This can ultimately facilitate the construction of high-performance FFT-based solvers for PDEs with general boundary conditions.", "description": "This contribution presents a package (FourierContinuation.jl) for building numerical continuation operators that enable discrete Fourier series representations of non-periodic functions while mitigating the Gibb's phenomenon. These operators can produce highly-accurate derivative approximations that are suitable for constructing high-order FFT-speed solvers for hyperbolic and parabolic PDEs governing a wide range of physical applications [1,2]. FC-based solvers have been demonstrated to incur minimal to no numerical pollution errors [1,2], making them ideal for fast and accurate long-time or long-distance simulation for scientific computing.\r\n\r\nMany underlying methodologies for numerically constructing such operators, which are often based on least squares problems, tend to be fundamentally ill-conditioned. Hence, implementation is non-trivial, requiring high-precision or symbolic arithmetic (for this task, Julia is well-suited). Here we provide a generalized tool for researchers to use the most common approaches for producing these periodic extensions of smooth functions, and to develop/test new ones.\r\n\r\n[1] A stable high-order FC-based methodology for hemodynamic wave propagation, F. Amlani and N.M. Pahlevan.\r\nJ. Comput. Phys. (2020)\r\nhttps://doi.org/10.1016/j.jcp.2019.109130\r\n\r\n[2] An FC-based spectral solver for elastodynamic problems in general three-dimensional domains, F. Amlani and O.P. Bruno.\r\nJ. Comput. Phys. (2016)\r\nhttps://doi.org/10.1016/j.jcp.2015.11.060", "recording_license": "", "do_not_record": false, "persons": [{"code": "MNNXKE", "name": "Ivan MENDEZ CRUZ", "avatar": null, "biography": null, "public_name": "Ivan MENDEZ CRUZ", "guid": "8976cef8-d818-5b18-aeae-93ad868cdbb0", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/MNNXKE/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/YFYL3K/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/YFYL3K/", "attachments": []}, {"guid": "f6d31f3f-1296-561f-8441-2acefdc8d98f", "code": "9TAF9T", "id": 70803, "logo": null, "date": "2025-10-02T11:21:00+02:00", "start": "11:21", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70803-unfoldsim-jl-a-package-for-simulating-event-based-time-series-data-for-eeg-and-beyond", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/9TAF9T/", "title": "UnfoldSim.jl: A package for simulating event-based time series data for EEG and beyond", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "UnfoldSim.jl is used to simulate multivariate time series, with a focus on continuous EEG data based on event-related potentials (ERPs). Its modular approach allows users to specify complex experimental designs, parameterized response functions, event order and temporal structure, and different noise specifications. Using abstract types and multiple dispatch, each ingredient can easily be exchanged and customized, allowing users to tailor the simulation to their needs.", "description": "The use cases for simulated (EEG) data are manifold. They include conducting power analyses, testing and comparing different analysis tools, validating statistical methods, creating example data for educational purposes, and illustrating conceptual issues.\r\n\r\nIn our work, we often analyze data containing (temporally) overlapping events (e.g., stimulus onset and button press, or consecutive eye-fixations), non-linear effects, and complex experimental designs. Therefore, we created a simulation package that can be used to simulate EEG data with these properties.\r\n\r\nHowever, the properties required for the simulated data vary by application, and therefore, a flexible and modular simulation approach is needed. For this reason, one of the fundamental design principles of UnfoldSim.jl is \u2018modularity\u2019, allowing users to adapt the simulation to their specific research question or application.\r\n\r\nFor a successful simulation, the user provides four ingredients:\r\n1.\tAn experimental design, with both categorical and continuous variables, and support for sequences of different event types (e.g. stimulus - response).\r\n2.\tEvent basis functions specified via linear or hierarchical models.\r\n3.\tAn inter-event onset distribution describing the temporal spacing of the events.\r\n4.\tA noise specification. \r\n\r\nMany of the simulation ingredients can also be nested to allow more complex functionalities as for example multi-channel simulations for EEG data.\r\n\r\nThis talk will give an overview of the building blocks of the package (e.g., the implementation of the simulation ingredients using abstract types) and guide you through a simple simulation example to illustrate the simulation workflow and showcase the influence of different parameters on the simulated data.", "recording_license": "", "do_not_record": false, "persons": [{"code": "KCEE3Z", "name": "Judith Schepers", "avatar": null, "biography": null, "public_name": "Judith Schepers", "guid": "dfb0f489-b035-5ed6-bb29-04af117f52d2", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/KCEE3Z/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/9TAF9T/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/9TAF9T/", "attachments": []}, {"guid": "28d55f46-6e7c-5e32-83dd-5dde73f31e46", "code": "C8HLP7", "id": 70541, "logo": null, "date": "2025-10-02T11:24:00+02:00", "start": "11:24", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70541-this-is-a-metaline-how-comments-can-add-parsable-meaning-to-code", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/C8HLP7/", "title": "#~ This is a metaline: How comments can add parsable meaning to code", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "The concept of **embedding formalized meaning into a Julia file's comments by means of brief, simple, expressive and extendable metadata** is firstly motivated at JuliaCon 2025 in Pittsburgh.\r\n\r\n**In Paris**, this talk will **greatly expand on the previous one, in depth and breadth** by illustrating the various benefits gained from adding parsable meaning to a file's content.\r\n\r\nTo this end, **`GoLiterate` will be formally introduced**. It serves a variety of different packages and facilitates interactions between them.", "description": "The primary aim of this talk is to motivate embedding formalized meaning into a Julia file's comments **by exhibiting a range of assorted examples.** The same principles are easily extended to other programming languages or project data in general.\r\n\r\nFor the purpose of illustration a new package will be introduced: `GoLiterate`. It is an extended adaptation of the well established `Literate` by *Fredrik Ekre et al.* Thanks \ud83c\udf37 The **handling of metadata is left to a purpose built back end** named `GoMeta`. Later it will be shown how prospective contributors may take advantage of its **modular design concept** in order to easily and fully integrate their own extensions. It is hoped that this approach will eventually yield a **pure Julia App serving an audience which extends beyond the Julia community**.\r\n\r\nBut first, what can actually be gained from embedding formalized metadata within comments? Let's consider **four, seemingly unrelated use cases** which together ought to resonate with the vast majority of JuliaCon's attendees:\r\n\r\n1. Over time a large number of Julia files can accumulate. Wouldn't it be wonderful to speedily retrieve any desired, long forgotten poignant example, powerful code snippet, insightful explanatory statement or laboriously tweaked plot related to some task at hand? Adding **meaning** to content can change the game. On top of that, an instrument solving this problem might prove even more invaluable when collaborating in a group sharing a repository of documents and it is also not restricted to Julia files only.\r\n\r\n2. One might want to (inter)link a certain fragment of one file with another fragment, possibly contained in some other file. E.g.: linking an intricate piece of code to an enlightening explanation or a clarifying toy example. Occasionally it could also be handy to swap entire sections of text / code of one file with alternate versions or update them centrally if used in multiple places, depending on the circumstances.\r\n\r\n3. Would you like to harness the power of `Documenter`, enjoying simplified control while sticking to pure Julia files rather than having to deal with additional markdown documents? More generally, a Julia file can serve as a source for all kinds of outputs.\r\n\r\n4. Being able to control how individual chunks of code / text are being executed, respectively rendered, can be beneficial. E.g.: certain passages of a file might be intended for personal usage only. Marking them as such would allow automated removal before sharing. Another example is the ability to restrict execution to code blocks pertaining to a particular group only.\r\n\r\nThe defining concept of `Literate` consists of storing chunks of code and text together in pure Julia files. These serve as source files which can subsequently be converted into a variety of formats such as notebooks, standard markdown or even documenter markdown files. A suitable system for metadata **allows exerting precise control over these format's wide ranging capabilities from within the Julia source file** and provides the means to generalize `Literate`'s core idea and more. The concept extends well beyond Julia and **can be applied not only to other programming languages but to all sorts of text documents** such as lectures notes or any number of scripts concerning a project, research or otherwise.\r\n\r\nSo what does the proposed schema for metadata look like?\r\n\r\nAny line starting with `#~` is considered to be a `metaline` i.e.: it will be processed as metadata. **Contiguous metalines** constitute a coherent `block of metadata` separating it from others. By **positioning such a block just above a chunk of code or text** [no vertical spacing in between] the former gets attached to the latter \u2013 similar to how docstrings get associated with Julia structures following them. For the purpose of `inheritance`, a **hierarchy of blocks of metadata** can be defined by prepending one or more `#` to the first metaline of a given block which works analogously to markdown's syntax for titles.\r\n\r\nThe current implementation of `GoLiterate` accepts `standard metalines` [more below] with its content consisting of key-value pairs where the keys are restricted to one word and the values may include general strings or vectors. Like this a `name` or `title` can be provided and `labels` can be applied. A `UID` is automatically generated and assigned to each block of metadata. This way a targeted search for particular sections of code or text becomes possible as well as interlinking various sections. \r\n\r\nThe execution of some tasks may be so common that they are deserving of a simplified syntax to facilitate their usage. Others may benefit or even necessitate custom tailored metadata to enable full functionality, as is the case for instructions passed to `Documenter`. The current schema resolves this by introducing `specialized metalines`. E.g.: a leading `#~d` instructs the parser to process the subsequent content for the usage with `Documenter`. Appending a `!` to a metaline declaration, `#~!`, causes its content to be **retained but not processed**.\r\n\r\nAll this is readily exemplified:\r\n```\r\n# Start file.jl\r\n\r\n#~  [Metadata Block 1]\r\n#~\r\n\r\n ... code/text\r\n\r\n##~ [Metadata Block 2]\r\n#~!\r\n#~\r\n\"\"\"\r\n\taddone(x)\r\nThis is a docstring for addone().\r\n\"\"\"\r\nfunction addone(x)\r\n\tx += 1\r\nend\r\n\r\n ... code/text\r\n\r\n#~   [Metadata Block 3]\r\n#~d jldoctest; output = false\r\naddone(2)\r\n\r\n# output\r\n\r\n3\r\n\r\n# End file.jl\r\n```\r\n\r\nAbove three distinct blocks of metadata are presented. The second of which can inherit attributes from the first. While the second block is attached to a chunk and thus gets applied to it, the first block has no direct effect on any code or text chunk. Its attributes are only passed on indirectly through inheritance.\r\n\r\nBlock 3 does not inherit anything from the blocks before since it is not a sub-block / sub-section of them. However, since it contains a *specialized metaline* meant for `Documenter`, the chunk of code following it will be fenced by ```` ```jldoctest; output = false \u2026 ``` ```` when the output format corresponds to `Documenter`'s markdown.\r\n\r\nWhich metadata gets invoked and how, can also be finely adjusted via function call [*user input*], *default settings*, *init file* etc. These methods of passing instructions can be used to conveniently and **efficiently inhibit any arbitrary directives issued by the metadata expressed in the file or override them altogether**; e.g.: effortlessly controlling which metadata gets displayed to prevent visual overload without editing the source file. This is achieved by permuting the precedences: ignore particular metadata and adhere to user input instead. This ability to prioritize and the hierarchical dependence between meta blocks reduces the need of lengthy, explicit written statements in the source file itself.\r\n\r\nThe **modular design of `GoMeta` enables future contributors to easily and fully integrate their own modules** by means of `PAnG!`: Firstly, the facilities for costume defined metadata get initiated [`P`lug]. Next, `GoMeta` must be able to parse and process [`A`bsorb] that new type of data contained in the file a`n`d, finally, act on it [`G`o`!`].\r\n\r\nTime permitting, this talk's various examples will demonstrate not just the above described features but many more which for the sake of brevity have been omitted from this already dense overview.\r\n\r\nIn the near future, AI will automate much of the labeling. Crucially, if this black box is confined within a formalized framework of metadata, it can become train- and tunable.", "recording_license": "", "do_not_record": false, "persons": [{"code": "QXHTX9", "name": "Jeroen Sieburgh", "avatar": "https://pretalx.com/media/avatars/QXHTX9_gVzGBYP.webp", "biography": null, "public_name": "Jeroen Sieburgh", "guid": "c40f5ee5-e1e7-50bf-b1f6-245dae2fec9f", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/QXHTX9/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/C8HLP7/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/C8HLP7/", "attachments": []}, {"guid": "1195d754-c51f-554a-ad17-da81707c02ad", "code": "LMDDXV", "id": 70936, "logo": null, "date": "2025-10-02T16:30:00+02:00", "start": "16:30", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70936-model-discovery-of-dynamics-using-universal-differential-equations-and-sciml", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/LMDDXV/", "title": "Model Discovery of Dynamics using Universal Differential Equations and SciML", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Model Discovery is a symbolic-numeric method that modifies the physical equations of a dynamic system in order to fit test or telemetry data. In this talk, we introduce the Model Discovery pipeline in the Julia product Dyad, which builds upon Universal Differential Equations and other SciML methods.", "description": "Dynamic model calibration to data is a challenging problem, sometimes due to the nonlinear and stiff nature of the dynamic problem, the number of parameters to calibrate, or the fact that model has too many simplifying assumptions. In the context of component-based dynamic models, where each component represents a physical object, different levels of fidelity are used to construct each component, which are then composed at the system level for calibration against data. So when model calibration fails, a concrete possibility is that the components themselves are low fidelity. \r\n\r\nIn this scenario, Model Discovery can be used to identify which components and equations to augment and the exact symbolic expressions to augment with. That is the subject of this talk. Concretely, model discovery consists of four steps: \r\n1. universal differential equations, which augments the dynamics with a neural network and regresses against the parameter of the neural network. Our implementation of this builds upon ModelingToolkitNeuralNets.jl.\r\n2. output masking, a heuristic which determines which equations are the most critical to change in order to fit the data. \r\n3. sensitivity analysis, built upon using SciMLSensitivity.jl, which identifies the correspondingly significant variables which influence the corrections. The above two steps reduce the input and output dimensions of the neural network. \r\n4. Symbolic regression, which builds upon SymbolicRegression.jl, which approximates the neural network with symbolic regression using important parameters and variables from the original model.  \r\n\r\nFinally, this talk will demonstrate the method on a practical example of a thermo-fluid system called a cargo box, which is a fundamental unit in the modeling of Refrigerated Transport Units.", "recording_license": "", "do_not_record": false, "persons": [{"code": "AZXJH3", "name": "Ranjan Anantharaman", "avatar": null, "biography": "Ranjan is a sales engineer at JuliaHub, where he works with JuliaHub's modeling and simulation customers. He has a PhD in Applied Mathematics from MIT.", "public_name": "Ranjan Anantharaman", "guid": "c9f62ce5-71bc-598a-ab2b-7291aa355b02", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/AZXJH3/"}, {"code": "HBZFM9", "name": "Sebastian Miclu\u021ba-C\u00e2mpeanu", "avatar": "https://pretalx.com/media/avatars/HBZFM9_fL6acie.webp", "biography": null, "public_name": "Sebastian Miclu\u021ba-C\u00e2mpeanu", "guid": "ffa7b5f6-4c9a-5588-8448-32faf175811d", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HBZFM9/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LMDDXV/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LMDDXV/", "attachments": []}, {"guid": "3a734e46-6bfc-5898-81ab-7dd3f9803202", "code": "UHSVUL", "id": 70024, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/UHSVUL/ket-jl-log_48SOPP8.png", "date": "2025-10-02T16:33:00+02:00", "start": "16:33", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70024-ket-jl-toolbox-for-quantum-information-nonlocality-and-entanglement", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/UHSVUL/", "title": "Ket.jl: Toolbox for quantum information, nonlocality, and entanglement", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Ket.jl is an open-source Julia package tailored for quantum information science, focusing on nonlocality and entanglement. It offers a suite of tools for constructing and processing quantum states, measurements, and channels, all designed to work seamlessly with JuMP. By supporting arbitrary-precision arithmetic, Ket.jl enables researchers to perform high-precision computations essential for exploring quantum phenomena and solving complex optimization problems. Ket.jl's design facilitates integration into optimization workflows, making it a valuable tool for both theoretical investigations and practical applications in quantum information science.", "description": "## Key Features\r\n\r\n### Quantum State and Measurement Construction\r\n- Generate mutually unbiased bases (MUBs) and symmetric informationally complete POVMs (SIC-POVMs) with arbitrary precision.\r\n- Create well-known quantum states, including Bell states, GHZ states, W states, Dicke states, and the super-singlet.\r\n- Produce uniformly distributed random states, unitaries, and POVMs for simulation and testing purposes.\r\n\r\n### Entanglement and Nonlocality Analysis\r\n- Compute entanglement measures such as entanglement entropy, entanglement robustness, and Schmidt number.\r\n- Evaluate local and Tsirelson bounds of Bell inequalities using parallelized algorithms.\r\n- Transform between different representations of Bell inequalities, including Collins-Gisin, probability, and correlation forms.\r\n\r\n### Optimization Integration\r\n- Leverage JuMP for defining and solving optimization problems involving quantum systems.\r\n- Utilize functions like `partial_trace`, `partial_transpose`, and `permute_systems` within optimization workflows.\r\n- Implement the DPS hierarchy constraints directly for semidefinite programming tasks.", "recording_license": "", "do_not_record": false, "persons": [{"code": "3AXFXH", "name": "S\u00e9bastien Designolle", "avatar": "https://pretalx.com/media/avatars/3AXFXH_dndyg6r.webp", "biography": "I'm an INRIA researcher based in **Lyon** (France).\r\nI'd like to meet people from the **OSCAR** project.\r\n\r\nMy research interests lie mostly at the interface between:\r\n - **quantum information**: Bell nonlocality, incompatibility of quantum measurements...\r\n - **convex optimisation**: semidefinite programming, Frank-Wolfe algorithms...", "public_name": "S\u00e9bastien Designolle", "guid": "bed0c65a-19f9-5806-b443-39056bb0129f", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/3AXFXH/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UHSVUL/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UHSVUL/", "attachments": []}, {"guid": "ed09d22f-6174-584b-9498-bd260136ba60", "code": "XPBVFA", "id": 72025, "logo": null, "date": "2025-10-02T16:36:00+02:00", "start": "16:36", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72025-improving-the-performance-of-gnns-using-sparse-linear-algebra", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/XPBVFA/", "title": "Improving the Performance of GNNs Using Sparse Linear Algebra", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Graph Neural Networks (GNNs) are powerful models for learning from graph-structured data, but their performance can depend critically on how message passing is implemented. In this talk, I will present my Google Summer of Code project with the Julia organization, aimed at accelerating GNNs by improving support for sparse computations on GPUs. By leveraging sparse-dense matrix multiplications\u2014rather than the traditional gather-scatter paradigm\u2014for layers such as graph convolutions, we can reduce memory overhead and improve performance. These enhancements are integrated into the GraphNeuralNetworks.jl package, helping bring Julia\u2019s GNN ecosystem closer to state-of-the-art frameworks in other languages.", "description": "Graph Neural Networks (GNNs) are increasingly used in domains such as recommendation systems, bioinformatics, and social network analysis. In Julia, the GraphNeuralNetworks.jl package provides a composable and elegant framework for experimenting with GNN architectures. However, performance limitations\u2014particularly on GPUs\u2014can hinder scalability in real-world applications.\r\n\r\nThis talk presents my Google Summer of Code project, which focuses on improving the performance of key GNNs layers by using sparse-dense matrix multiplication instead of the classical gather-scatter message-passing scheme. The gather-scatter approach often creates large intermediate tensors, whereas sparse matrix operations can fuse computation and aggregation more efficiently. These optimizations are especially relevant for GPU acceleration, where memory overhead is a significant concern.\r\nThe project involves integrating these performance improvements into GraphNeuralNetworks.jl, benchmarking them, and making them accessible to the broader Julia community. While the project builds on existing CUDA.jl capabilities, it also explores how to best utilize them in practice for GNNs workloads.\r\n\r\nThis talk will cover the following topics:\r\n- GNNs and efficient message-passing strategies\r\n- Integration of sparse GPU operations into GraphNeuralNetworks.jl\r\n- Benchmarking insights and future directions", "recording_license": "", "do_not_record": false, "persons": [{"code": "HNYVDL", "name": "Davide Ferre'", "avatar": "https://pretalx.com/media/avatars/HNYVDL_at3GeqP.webp", "biography": "I'm a PhD student in Computer Science at Universit\u00e9 C\u00f4te d\u2019Azur, working within the COATI team, a joint group between Inria and CNRS. My research focuses on machine learning and graph theory, and I enjoy working at the intersection of theory and practice. As a 2025 Google Summer of Code contributor with the Julia organization, I'm working on integrating GPU-accelerated sparse operations into the GraphNeuralNetworks.jl package to support more efficient implementations of graph neural network layers.", "public_name": "Davide Ferre'", "guid": "2d6140c5-1657-533d-ada2-2f27389ccb7b", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HNYVDL/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/XPBVFA/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/XPBVFA/", "attachments": []}, {"guid": "eacad880-b70e-5464-8d1c-e890a68bb579", "code": "TUFEHD", "id": 72318, "logo": null, "date": "2025-10-02T16:39:00+02:00", "start": "16:39", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72318-how-can-we-design-julia-for-the-next-generation", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/TUFEHD/", "title": "How can we design Julia for the next generation?", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "You and I both think Julia is **the best** language! Faster than Python, more flexible than R, better syntax than C++. So why are these other languages still dominant in scientific computing, and growing? Why is it a \"weird choice\" to teach your class in Julia? \r\n\r\nThis talk explores what Julia could look like if we designed it _entirely for future users_. Drawing from the Pluto project's work with students and teachers, plus lessons from other language ecosystems, I'll share a vision for making Julia's technical advantage actually accessible to the people who could benefit most from it.", "description": "If Julia is so awesome... (it is!) then why is adoption in the scientific community not exploding? Sure, Python and R will be popular tomorrow simply because they are popular today. But there's more than that! Accessibility is key! What if we work on Julia with the belief that it will be picked up by more scientists five years from now? What would we work on next?\r\n\r\nWorking on the [Pluto project](https://plutojl.org), we focus on students, teachers and new Julia users. We see how they benefit from using Julia, but we also see the hurdles when getting started with the language. Julia can be hard to learn, especially when your background is not in mathematics. We follow Julia's development closely, and it can be frustrating to see so few core projects to tackle this. It feels like we're optimising for expert users while the next generation of scientists is learning Python instead.\r\n\r\nI would like to share a vision of what Julia _could_ look like in 5 years, if our focus was _entirely_ on future users. We could have a version of Julia that any scientist could pick up \u2013 code that stays reproducible by default, documentation that welcomes different backgrounds, AI-assisted learning tools and first-class developer tooling. \r\n\r\nJulia is the solution \u2013 the best language for our world's future scientists. Let's make sure they know how to use it.", "recording_license": "", "do_not_record": false, "persons": [{"code": "ECC83Q", "name": "Fons van der Plas", "avatar": "https://pretalx.com/media/avatars/ECC83Q_5CxYWbQ.webp", "biography": "\ud83c\udf3c Main developer of [Pluto.jl](https://plutojl.org/), currently working at TU Eindhoven to develop the course [Bayesian Machine Learning](https://github.com/bmlip/colorized).\r\n\r\nCome talk to me at the conference! I would love to learn more about what you are working on :) If you are teaching using Julia, definitely come talk to me!", "public_name": "Fons van der Plas", "guid": "17bd1580-e1fc-5230-9599-f2cc7f6c0695", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/ECC83Q/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/TUFEHD/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/TUFEHD/", "attachments": []}, {"guid": "7d376a97-a20d-525c-b5a3-46326dd0e5e1", "code": "7WJ3BR", "id": 72291, "logo": null, "date": "2025-10-02T16:42:00+02:00", "start": "16:42", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72291-from-resonator-networks-to-markov-chains-studying-learning-metamaterials-using-the-julia-ecosystem", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/7WJ3BR/", "title": "From resonator networks to Markov chains: Studying learning metamaterials using the Julia ecosystem", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "With the ever increasing computational cost of machine learning applications in mind, the need to explore alternative learning paradigms becomes apparent. One such paradigm would be to exploit the massive parallelism of physical systems. We introduce a metamaterial based on coupled nonlinear resonators that learns in-situ, and show how we are able to study it seamlessly across various levels of abstraction, building on Julia and the extensive ecosystem for scientific computing.", "description": "There exist many physical systems that show learning behaviour in response to external stimuli spanning large ranges of complexity, from slime moulds to the human brain. These systems are generally able to learn complex tasks at a fraction of the energy required by modern machine learning algorithms, simply by following a trajectory imposed by physical laws. As such, investigating the constraints and possibilities that arise from these systems existing in a physical world, compared to their in-silico counterparts which are (only) limited by the properties of the hardware they are implemented on, may open an avenue to new and highly energy-efficient computing and learning paradigms.\r\n\r\nWe introduce a system of two-dimensional, coupled nonlinear resonators, exploiting naturally occurring degeneracies to store short- and long-term memory, the two main ingredients required for learning, in its physical dynamics. By carefully choosing the couplings between the resonators, we are able to create a feedback loop between short- and long-term memory, which causes the system to learn.\r\n\r\nIn this poster, we show how we use Julia to bring together many different packages from different corners of the ecosystem to study and optimise the learning dynamics of this metamaterial across many levels of abstraction, from stochastic differential equations over dynamical systems theory to Markov chains. We argue that this ability to easily look at the problem from many perspectives, combined with Julia's general speed is highly beneficial to computationally study systems that might not be optimised to run efficiently on classical computers, but in the real, physical world.", "recording_license": "", "do_not_record": false, "persons": [{"code": "DXYQBD", "name": "Henrik Wolf", "avatar": "https://pretalx.com/media/avatars/DXYQBD_4DPnZEG.webp", "biography": "I was trained as a theoretical physicist and complex systems researcher, with a heavy focus on interdisciplinary research. After working mostly on the complexity of human mobility, I now study ways in which we can use physics to build learning metamaterials.", "public_name": "Henrik Wolf", "guid": "8a3a20aa-10b6-514a-bc7d-c6be1c38f889", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/DXYQBD/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7WJ3BR/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/7WJ3BR/", "attachments": []}, {"guid": "41fc48f8-416a-54ef-8747-fca63320cfb8", "code": "VTRDZ8", "id": 70893, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/VTRDZ8/Screenshot_ofbBWIs.png", "date": "2025-10-02T16:45:00+02:00", "start": "16:45", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70893-end-to-end-parameter-learning-and-applications-to-soil-science", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/VTRDZ8/", "title": "End-to-End Parameter Learning and applications to soil science.", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "In this presentation, we showcase Parameter-Learning in soil carbon modeling, a scientific machine learning paradigm where a process-based model is parameterized through the use of a neural network. This enables us to learn latent soil properties from soil carbon data, leveraging the Julia\u2019s ML and auto-differentiation ecosystems. The resulting hybrid model is more explainable through the latent variables, and more robust as it is constrained to a process-based model.", "description": "Soils are a crucial component of the global carbon cycle, and store more carbon than the atmosphere and vegetation combined. Therefore learning latent soil properties (such as microbial carbon use efficiency or sorption capacity) is important for understanding the fate of soil carbon under climate change.\r\n\r\nHowever, ground soil sampling data is sparse, and only soil carbon data is available with a sufficiently comprehensive global distribution. Latent soil properties can nevertheless be inferred by parameterizing a process-based model that predicts soil carbon, where the parameters of the process-based model are our latent variables of interest. Previously, parameterizing these models for global predictions was a two-step process, first a data assimilation process, followed by training the ML model on the results of the data assimilation. Thanks to the recent progress of auto-differentiation frameworks, a one-step end-to-end approach is now possible; a neural network predicts parameters to be fed to a process-based model, and its weights are optimized through auto-differentiation to give fast, reliable, scalable results.\r\n\r\nContributors: Bernhard Ahrens, Leo Rossdeutscher, Marco Paina, Thomas Wutlzer, Marion Schrumpf, Markus Reichstein", "recording_license": "", "do_not_record": false, "persons": [{"code": "8M9HTH", "name": "M. Ayoub Chettouh", "avatar": null, "biography": "First year PhD Researcher at the Max Planck Institute for Biogeochemistry with a background in theoretical physics. Julia user since 2017. \r\n\r\nResearch Summary: Parameterizing physical and other models using neural networks, applications to soil science.\r\n\r\nhttps://github.com/Qfl3x", "public_name": "M. Ayoub Chettouh", "guid": "cbf44c60-f7e4-5d3c-8f7a-5fa65b83aadd", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/8M9HTH/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VTRDZ8/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VTRDZ8/", "attachments": []}, {"guid": "40d96156-7cd5-5eeb-a5b3-0845a1c7e016", "code": "9HPK7G", "id": 72193, "logo": null, "date": "2025-10-02T16:48:00+02:00", "start": "16:48", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72193-efficient-computation-of-quasi-periodic-green-s-functions-for-the-helmholtz-equation-with-qpgreen-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/9HPK7G/", "title": "Efficient computation of quasi-periodic Green's Functions for the Helmholtz Equation with QPGreen.jl", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "The Helmholtz equation is essential to modeling wave scattering problems in periodic structures. Green\u2019s functions, which solve linear partial differential equations with Dirac delta function sources \u03b4(x), provide a foundation for recasting these problems into boundary integral equations. While closed-form expressions exist for many settings, the quasi-periodic Green\u2019s function for the Helmholtz equation poses a unique challenge: it is defined as an infinite series that converges slowly as x2 \u2192 0 in 2D (or x3 \u2192 0 in 3D). \r\n\r\nIn this talk, we introduce\u00a0QPGreen.jl, which implements an FFT-based algorithm to efficiently evaluate these Green\u2019s functions. It combines spectral truncation and interpolation to overcome the limitations of direct series summation. Leveraging Julia\u2019s composability, we combine\u00a0QPGreen.jl\u00a0with\u00a0Inti.jl\u2014a boundary integral equation solver\u2014and demonstrate applications to wave scattering problems with quasi-periodic incident fields.", "description": "The integral equation method is a popular method for simulating acoustic wave scattering in periodic structures, requiring efficient evaluation of quasi-periodic Green\u2019s functions.\r\n\r\nFor the Helmholtz equation, the series representation of quasi-periodic Green\u2019s functions converges too slowly to be practical in boundary integral equation methods. To address this, we introduce QPGreen.jl (https://github.com/gregoirepourtier/QPGreen.jl), which implements an FFT-based algorithm based on [1]. The method proceeds in two stages:\r\n- Preparation step: precomputes various Fourier coefficients and evaluates the Green\u2019s function at grid points. This step is performed once per set of quasi-periodicities and wavenumbers.\r\n- Evaluation step: efficiently interpolates values between grid points to compute values at arbitrary points between precomputed grid nodes, delivering efficient and accurate pointwise evaluations.\r\n\r\nThis approach overcome the limitations of direct series summation while maintaining high accuracy, making it ideal when a large number of evaluations is required.\r\n\r\nBy leveraging Julia\u2019s composability and package extension capabilities, we combine QPGreen.jl with Inti.jl (https://github.com/IntegralEquations/Inti.jl) - a boundary integral equation solver - to efficiently solve 2D and 3D scattering problems involving periodic gratings.\r\n\r\n[1] B. Zhang and R. Zhang. An FFT-based algorithm for efficient computation of Green\u2019s functions for the Helmholtz and Maxwell\u2019s equations in periodic domains. SIAM Journal on Scientific Computing 40, B915 \u2013 B941 (2018).", "recording_license": "", "do_not_record": false, "persons": [{"code": "XPCZJR", "name": "Gr\u00e9goire Pourtier", "avatar": "https://pretalx.com/media/avatars/XPCZJR_Q2yYIo1.webp", "biography": "I am a PhD student in the research group for analysis and applications at TU Berlin. My research focuses on the numerical solution of scattering problems, primarily using boundary integral equation methods.", "public_name": "Gr\u00e9goire Pourtier", "guid": "f9956f7e-848d-5fbe-8703-3151485afd43", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/XPCZJR/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/9HPK7G/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/9HPK7G/", "attachments": []}, {"guid": "a93bb40e-fa0d-5073-ab14-59f7e71b6f12", "code": "RHX7CG", "id": 70946, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/RHX7CG/MedEye3D.j_usbP4Xl.png", "date": "2025-10-02T16:51:00+02:00", "start": "16:51", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70946-a-digital-twin-approach-for-advanced-supervoxel-visualization-for-multi-image-view-in-medical-imaging", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/RHX7CG/", "title": "A Digital Twin approach for Advanced Supervoxel Visualization for Multi-Image View in Medical Imaging", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "MedEye3d.jl is a vital Julia package for 3D medical image visualization, playing a crucial role in workflows from registration to segmentation analysis. This talk introduces recent advancements in MedEye3d.jl, stemming from GSoC, that push it towards a \"digital twin\" for medical imaging. We'll explore the implementation of a novel multi-image display mode, enabling concurrent visualization of registered heterogeneous studies (e.g., CT and PET). A key focus will be the introduction of dynamic supervoxel border rendering and interactive highlighting across these linked views, allowing for intuitive comparison and analysis. We'll also touch upon new features for interactive supervoxel association correction and control point annotations, significantly enhancing the tool's utility for registration, orientation, and detailed visual inspection. These enhancements leverage Julia's performance and ModernGL.jl for a responsive and powerful user experience, further solidifying Julia's position in the medical imaging domain.", "description": "The JuliaHealth ecosystem provides a powerful suite of tools for medical imaging, and MedEye3d.jl stands as a cornerstone for visual inspection throughout the processing pipeline. This presentation will detail the significant enhancements made to MedEye3d.jl during Google Summer of Code, transforming it into a more powerful and interactive \"digital twin\" for researchers and clinicians.\r\n\r\nMotivation: the need for richer, comparative visualization capabilities, particularly for multi-modal datasets and superpixel/supervoxel-based analyses.\r\n\r\nThe core of the talk will cover the key new features:\r\n- Advanced Multi-Image Supervoxel Visualization:\r\nDemonstration of the new side-by-side panel view for different imaging modalities (e.g., CT alongside PET), a step up from previous same-modality views.\r\nTechnical insights into dynamic texture creation and fragment shader management for heterogeneous studies using ModernGL.jl.\r\nShowcasing dynamically updated supervoxel borders that correspond to the current slice in both linked image views, with synchronized scrolling.\r\nHighlighting corresponding supervoxels across images based on cursor position, facilitating direct anatomical comparison.\r\n- Interactive Supervoxel Associations & Control:\r\nIntroduction of user interactions (via GLFW callbacks) to manually correct or confirm supervoxel associations between the displayed images.\r\nDiscussing the mechanism for users to flag incorrect default mappings and how MedEye3d.jl updates these associations, improving the tool's reliability through a user feedback loop.\r\n- Enhanced Annotation: Control Points for Registration & Orientation:\r\nImplementing control point annotations in multi-image mode, allowing users to mark corresponding anatomical landmarks across studies.\r\nDetailing the persistence of these annotations, including support for exporting them (e.g., in NIfTI or JSON sidecar files) to aid subsequent image registration processes or for maintaining orientation context.\r\nThroughout the presentation, we will emphasize how these features contribute to a more seamless and intuitive visual inspection process, crucial for detecting anomalies, verifying registration accuracy, and understanding complex spatial relationships in 3D medical data. We will also briefly touch upon efforts to improve package health, including documentation and testing, to foster community growth around MedEye3d.jl. The talk aims to showcase the practical power of Julia for developing sophisticated scientific visualization tools.", "recording_license": "", "do_not_record": false, "persons": [{"code": "KBBDXE", "name": "Divyansh Goyal", "avatar": "https://pretalx.com/media/avatars/KBBDXE_pPpO2OH.webp", "biography": "I engage with scientific computing and computational biology through Julia lang development, with applications in High-Energy Physics, Cryo-ET, and HPC. This includes two Google Summer of Code projects with The Julia Language (2024 & 2025). I am also an O'Reilly DEIJ Scholarship Recipient and a Microsoft Learn Student Ambassador.", "public_name": "Divyansh Goyal", "guid": "4ec4117c-d95b-5522-9082-81af19fda917", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/KBBDXE/"}], "links": [{"title": "MedEye3d.jl Package soure", "url": "https://github.com/juliahealth/MedEye3d.jl", "type": "related"}], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RHX7CG/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RHX7CG/", "attachments": []}]}}, {"index": 2, "date": "2025-10-03", "day_start": "2025-10-03T04:00:00+02:00", "day_end": "2025-10-04T03:59:00+02:00", "rooms": {"Jean-Baptiste Say Amphitheater": [{"guid": "68af52fe-2dd7-5711-b291-cb44e3f7a866", "code": "PGXWMP", "id": 79776, "logo": null, "date": "2025-10-03T08:30:00+02:00", "start": "08:30", "duration": "01:00", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79776-highs-the-story-so-far", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/PGXWMP/", "title": "HiGHS: The Story So Far", "subtitle": "", "track": "Optimization", "type": "Plenary", "language": "en", "abstract": "This talk presents HiGHS, the world's leading open-source solver for large-scale, sparse linear programming and mixed integer programming problems. We will discuss the types of problems handled by HiGHS, its solution methods, and some of the challenges we have faced in transforming them into software. We will also talk about scaling software, platforms, interfaces and maintaining the reliability of a widely used open-source project.", "description": "Ivet Galabova is the Development and Integration Manager of HiGHS, the world's leading open-source optimization software. Ivet is a mathematician and research software engineer with a strong interest in linear (LP), mixed integer (MIP) and quadratic (QP) programming problems, build systems and interfaces. Along with completing a PhD in Optimization and Operations Research at the University of Edinburgh, Ivet has been closely involved in the development of the HiGHS from the start of the project in 2017. HiGHS can be used via the HiGHS.jl package in Julia.", "recording_license": "", "do_not_record": false, "persons": [{"code": "JZHGRP", "name": "Ivet Galabova", "avatar": null, "biography": "Ivet Galabova is the Development and Integration Manager of HiGHS, the world's leading open-source optimization software. Ivet is a mathematician and research software engineer with a strong interest in linear (LP), mixed integer (MIP) and quadratic (QP) programming problems, build systems and interfaces. Along with completing a PhD in Optimization and Operations Research at the University of Edinburgh, Ivet has been closely involved in the development of the HiGHS from the start of the project in 2017. HiGHS can be used via the HiGHS.jl package in Julia.", "public_name": "Ivet Galabova", "guid": "93d72449-1131-5e6f-bd43-42a51d522393", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/JZHGRP/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/PGXWMP/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/PGXWMP/", "attachments": []}, {"guid": "41ea52e6-8e4e-55db-a171-13118a61a95e", "code": "ZUYGVA", "id": 67833, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/ZUYGVA/ortools_lo_dCOCxny.png", "date": "2025-10-03T09:30:00+02:00", "start": "09:30", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-67833-ortools-jl-access-google-s-solvers-through-jump", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZUYGVA/", "title": "ORTools.jl: access Google's solvers through JuMP", "subtitle": "", "track": "Optimization", "type": "Talk", "language": "en", "abstract": "In Julia, JuMP is the go-to modelling package for mathematical optimisation. As of this writing, Google's award-winning solvers have not been accessible through JuMP; which offers Julia's ease of use. ORTools.jl is changing this. Julia users will now have access to Google's Glop, CP-SAT, and PDLP solvers through JuMP as provided by the ORTools.jl package.\r\nThis talk offers an introduction to the features of the package and an overview of the difficulties we encountered.", "description": "ORTools.jl is a package that exposes Google\u2019s optimisation solvers to the Julia ecosystem. It achieves this goal through MathOptInterface, for JuMP and Optimization.jl compatibility, and doing necessary data structure conversions, currently primarily through Protocol Buffer, to target Google\u2019s MathOpt library.\r\n\r\nIt will expose Google\u2019s Glop, CP-SAT, and PLDP open-source solvers to the Julia community, giving access, for the very first time, to Google\u2019s solvers through JuMP as they would other solvers. In particular, [CP-SAT has won gold medals at the MiniZinc challenge for more than a decade](https://www.minizinc.org/challenge/) and [PDLP has sparked new interest in first-order methods for linear programming at extremely large scale](https://arxiv.org/abs/2501.07018). \r\n\r\nWriting this package was a technical challenge, as the Google solvers are only accessible through a C++ interface, a language that is notoriously hard for interop. The first version of ORTools.jl uses a Protocol Buffer interface and a very small C API to call the solvers. It relies on a JLL package to build OR-Tools, ensuring a smooth experience for end users", "recording_license": "", "do_not_record": false, "persons": [{"code": "EVRZ99", "name": "Thibaut Cuvelier", "avatar": "https://pretalx.com/media/avatars/EVRZ99_wy2I3DL.webp", "biography": "Thibaut Cuvelier is currently a software engineer at Google Research, in the Operations Research team. He received a PhD in telecommunications from CentraleSup\u00e9lec (universit\u00e9 Paris-Saclay, France). He is currently working on applications of operations research and reinforcement learning in logistics.", "public_name": "Thibaut Cuvelier", "guid": "fe6140ad-8a8c-5e72-96e0-c8b665999d27", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/EVRZ99/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZUYGVA/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZUYGVA/", "attachments": []}, {"guid": "c65ee9ac-7e0c-55c6-a5ae-4c9cbd5d71b5", "code": "UMRDXY", "id": 70209, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/UMRDXY/CrystalNet_OC0uZmi.jpg", "date": "2025-10-03T10:00:00+02:00", "start": "10:00", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70209-crystalnets-jl-efficient-periodic-graph-canonization-made-accessible-for-the-scientific-community", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/UMRDXY/", "title": "CrystalNets.jl: efficient periodic graph canonization made accessible for the scientific community", "subtitle": "", "track": "Optimization", "type": "Lightning talk", "language": "en", "abstract": "CrystalNets.jl is a library dedicated to the topological identification of crystalline materials, a seminal problem in materials science. This requires solving the periodic graph canonization problem, a superset of the well-known graph isomorphism problem, but also several other issues ranging from bond detection in chemical structures to the design of an optimized sparse rational system solver. CrystalNets.jl boasts a unique combination of crystal structures coverage and order-of-magnitude performance improvements compared to other topological analysis tools. We will discuss the design of the library, as well as the development of its web interface which allows an easy access to the performance of the Julia code, even for non-programmers in the scientific community.", "description": "Crystals form a very large class of materials, whose structural organization consists in an atomic pattern periodically repeated across the three dimensions of space. The topology of a material is the information that can be derived from the graph of its chemical bonds: in the case of crystals, this graph is thus periodic. The study of the topology of materials is a crucial matter in materials science, since it is a primary mean of classification, and families of structures sharing a topology can exhibit similar properties. Yet, identifying the topology of a crystalline material is computationally difficult, since it requires solving the problem of periodic graph canonization.\r\nGraph canonization consists in finding, for a given graph, a unique \"canonical\" representation; in some sense, its \"name\". It is a harder problem than graph isomorphism, which consists in checking whether two graphs are identical up to the renaming of their vertices, since the the latter can be solved by computing the canonical representations of the two graphs and checking whether they are identical. Yet, it is unknown whether the simpler problem of graph isomorphism can be solved in polynomial time.\r\n\r\nCrystalNets.jl is a Julia library that implements a full topology identification toolkit, taking as an input the chemical representation of a material and outputting the name of its topology if it is known in the scientific literature, or its canonical representation otherwise. It thus fully solves the periodic graph canonization problem, but also encompasses many options for the extraction of the graph from the chemical input. Its algorithm is polynomial in the vast majority of real crystal cases, but falls back to a slower strategy when needed to ensure coverage of all possible structures. Its can analyze databases of millions of structures easily, significantly outperforming similar tools in both accuracy and speed.\r\nWe will discuss the design and implementation of the library and some related tools for the study of periodic graphs.\r\n\r\nIn order to make it more accessible for the scientific community, a web interface was developed to allow an intuitive access to the powerful algorithms of CrystalNets.jl for anyone. We will also discuss the implementation of this interface, and the efforts to make the performance of this Julia program reach a wider audience.\r\n\r\nGitHub repository: https://github.com/coudertlab/CrystalNets.jl\r\nDocumentation: https://coudertlab.github.io/CrystalNets.jl/dev\r\nWeb interface: https://progs.coudert.name/topology", "recording_license": "", "do_not_record": false, "persons": [{"code": "D9JTTJ", "name": "Lionel Zoubritzky", "avatar": null, "biography": "Post-doctorate researcher at Coll\u00e8ge de France in computational chemistry, and Julia teacher at \u00c9cole des Mines - PSL.", "public_name": "Lionel Zoubritzky", "guid": "63787446-8a4e-506f-9fc2-589289e3d5b7", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/D9JTTJ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UMRDXY/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/UMRDXY/", "attachments": []}, {"guid": "bdc028d5-04af-5b82-a78f-a74e76d25b13", "code": "TZ9FBJ", "id": 70630, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/TZ9FBJ/graphlab_cLnuizw.png", "date": "2025-10-03T10:10:00+02:00", "start": "10:10", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70630-graphlab-jl-a-julia-framework-for-graph-partitioning-algorithms", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/TZ9FBJ/", "title": "GraphLab.jl: A Julia Framework for Graph Partitioning Algorithms", "subtitle": "", "track": "Optimization", "type": "Lightning talk", "language": "en", "abstract": "We present GraphLab.jl, a Julia package designed to facilitate the study, experimentation, and research of graph partitioning. GraphLab.jl provides a framework for exploring the principles and trade-offs of partitioning algorithms through hands-on tools. It implements a growing set of methods\u2014including coordinate, inertial, and spectral bisection, random spheres, space-filling curves, and nested dissection\u2014with support for recursive partitioning. The package includes routines for generating adjacency matrices, computing partition quality metrics, benchmarking problems, and visualizing partitioned graphs. GraphLab.jl enables integration with external graph partitioning software, thus allowing users to compare additional methods and results in a unified environment. Our work aims to introduce Julia's capabilities to learners and researchers engaging in graph theory and related partitioning problems.", "description": "Graph partitioning is a fundamental problem with wide-ranging applications in computational biology, social network analysis, high-performance computing, and distributed systems. Partitioning large graphs into loosely connected subsets of roughly equal size promotes parallel execution, reduces communication overhead, and provides insights into the structure of complex networks.\r\n\r\nOver the years, a plethora of graph partitioning techniques have been developed. Evaluating their effectiveness and comparing their trade-offs are key to building intuition and practical understanding. We address this need with GraphLab.jl, a package designed to facilitate the study, experimentation, and research of graph partitioning in an integrated, hands-on workflow. GraphLab.jl offers an accessible framework that enables users to experiment with algorithms, visualize results, and assess partition quality in an interactive environment. The package implements a diverse set of partitioning algorithms, including coordinate, inertial, and spectral bisection, random spheres, space-filling curves, and nested dissection. These methods can be applied recursively for hierarchical partitioning or multi-level strategies. GraphLab.jl also provides routines for generating adjacency matrices, computing partition quality metrics, benchmarking problems, and visualizing partitioned graphs. It allows integration with external graph partitioning software, thus enabling users to compare additional methods and results in a unified environment.\r\n\r\nOur work aims to contribute to the educational efforts of the Julia graph community and to assist in introducing Julia's capabilities to learners and researchers engaging in graph theory and related partitioning problems.\r\n\r\nThe code is openly available on GitHub: https://github.com/lechekhabm/GraphLab.jl", "recording_license": "", "do_not_record": false, "persons": [{"code": "J89QDK", "name": "Dimosthenis Pasadakis", "avatar": "https://pretalx.com/media/avatars/J89QDK_fLzzjPq.webp", "biography": "I am a postdoctoral fellow at Universit\u00e0 della Svizzera italiana (USI) in Lugano, Switzerland. The focus of my research is centered around algorithms for graph learning and combinatorial optimization for graph clustering and anomaly detection. Currently, I am leading the project \u201cDirected acyclic graph partitioning for scheduling tasks\u201c, financed by the Huawei Research Center Z\u00fcrich. I have an MSc and PhD in Computational Science from USI, and a Degree of Physics from the Aristotle University of Thessaloniki, Greece.\u2028\r\n\r\nPersonal website: https://dmspas.github.io/", "public_name": "Dimosthenis Pasadakis", "guid": "05115707-8425-56b7-b040-37bcb205cedb", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/J89QDK/"}, {"code": "AEP8AJ", "name": "Malik Lechekhab", "avatar": "https://pretalx.com/media/avatars/AEP8AJ_yX52KJ5.webp", "biography": "I'm a PhD student in High Performance Computing at the Advanced Computing Laboratory, Universit\u00e0 della Svizzera italiana. I hold a master\u2019s degree in Finance from HEC, University of Lausanne. My research interests include graph theory, anomaly detection, and computational finance. In my free time, I enjoy all mountain sports, from hiking to paragliding.", "public_name": "Malik Lechekhab", "guid": "bc99f85b-ba8e-5aa5-b7dc-07bd5326c740", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/AEP8AJ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/TZ9FBJ/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/TZ9FBJ/", "attachments": []}, {"guid": "06c6f476-41ce-5c05-b35a-3f204bbe7bc4", "code": "LC9S8P", "id": 72293, "logo": null, "date": "2025-10-03T10:20:00+02:00", "start": "10:20", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-72293-pace-jl-certified-solving-of-polynomial-systems-for-engineering-applications", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/LC9S8P/", "title": "PACE.jl: Certified Solving of Polynomial Systems for Engineering Applications", "subtitle": "", "track": "Optimization", "type": "Lightning talk", "language": "en", "abstract": "PACE.jl is a Julia package designed to address challenging problems in control theory and robotics through certified polynomial system solving. By combining symbolic elimination techniques with multi-precision interval computations, PACE.jl provides robust, certified solutions and an efficient alternative to traditional numerical methods. This talk will demonstrate its usage and performance through classic examples from control theory and robotics, while remaining accessible to anyone interested in certified polynomial system solving.", "description": "Engineering applications demand computational tools that provide both efficiency and mathematical rigor. While numerical methods offer speed, they often lack guarantees, and symbolic methods, while exact, can be computationally expensive. PACE.jl addresses this challenge by providing a unified framework that combines the best of both worlds, implemented in both C and Julia. \r\n\r\nIt combines symbolic elimination techniques, including (among others) discriminant varieties and Rational Univariate Representation, with numerical refinement using Newton's method and multi-precision interval computations.  In particular, PACE.jl brings together several specialized packages, some developed by our team and others by collaborators:\r\n\r\n- DiscriminantVariety.jl: Discriminant variety computations in polynomial systems [3]. \r\n- LACE.jl: Julia wrapper around a C library for Newton's method with interval arithmetic [4].\r\n- MPFI.jl: Wrapper for the C library MPFI for multiple-precision interval computations in Julia. See [5] for specifications. \r\n- RationalUnivariateRepresentation.jl: Algorithm from [1] for the Rational Univariate Representation (RUR) computation of zero-dimensional polynomial systems. \r\n- RS.jl: Julia wrapper around a C library that computes isolating intervals for the real roots of a polynomial or a zero-dimensional polynomial system. The package implements the algorithms from [2] and [6]. \r\n\r\nDuring this talk, we will demonstrate how to use PACE.jl to solve polynomial systems with certified guarantees in Julia, leveraging symbolic-numeric methods. As concrete examples, we will use established certified methods from control theory, including parameter identification in dynamical systems, stability analysis, and H-infinity norm computation, showing how PACE.jl handles both parameter-dependent and parameter-free systems. Our experiments show that PACE.jl achieves better performance than Maple's certified solvers, while remaining free and open-source.\r\n\r\nWhile our primary focus is on control theory applications, we will briefly mention PaceRobots.jl, a specialized package that applies PACE.jl to certify the kinematics of robotic systems, including Spherical Parallel Manipulators and Hexapod Robots. Due to time constraints, we won't be able to demonstrate PaceRobots.jl in detail, but this will give attendees a glimpse of the broader applicability of our methods.\r\n\r\nThe presentation will be accessible to practitioners across various fields, as polynomial systems appear naturally in diverse scientific problems.\r\n\r\nAll examples and code will be made available through our documentation and notebooks, which we plan to polish and enhance by the time of the talk:\r\n- PACE.jl: https://pace.gitlabpages.inria.fr/pace.jl/\r\n- Control Applications: https://pace.gitlabpages.inria.fr/pace.jl/applications-control/\r\n- Robotics Applications (PaceRobots.jl): https://pace.gitlabpages.inria.fr/robotics/pacerobots.jl/\r\n\r\n\r\nReferences:\r\n[1] Demin A., Rouillier F., and Ruiz J. (2024). Reading Rational Univariate Representations on lexicographic Groebner bases, preprint available at https://arxiv.org/abs/2402.07141\r\n[2] Kobel, A., Rouillier, F., and Sagraloff, M. (2016). Computing Real Roots of Real Polynomials ... and now For Real!, In ISSAC '16 Proceedings of the ACM on International Symposium on Symbolic and Algebraic Computation\r\n[3] Lazard and Rouillier (2007). Solving Parametric Polynomial Systems, Journal of Symbolic Computation\r\n[4] Revol, N. (2001). Interval Newton Iteration in Multiple Precision for the Univariate Case, research report available at https://inria.hal.science/inria-00072253v1/file/RR-4334.pdf\r\n[5] Revol, N. and Rouillier, F. (2005). Motivations for an arbitrary precision interval arithmetic and the MPFI library. Reliable Computing, 11(4), 275\u2013290\r\n[6] Rouillier, F. and Zimmermann, P. (2004). Efficient isolation of polynomial's real roots, Journal of Computational and Applied Mathematics, 162(1), 33-50", "recording_license": "", "do_not_record": false, "persons": [{"code": "3NJUPT", "name": "Christina Katsamaki", "avatar": "https://pretalx.com/media/avatars/3NJUPT_YvNFnjD.webp", "biography": "Currently a research engineer at Inria Paris developing symbolic computation software. I hold a PhD in Mathematics from Sorbonne Universit\u00e9 (Paris 6).", "public_name": "Christina Katsamaki", "guid": "4c119d58-6f10-5eb0-bed1-825c32bce910", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/3NJUPT/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LC9S8P/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LC9S8P/", "attachments": []}, {"guid": "a97680b7-8125-5fb2-9a3f-070889f8a0e5", "code": "LEVACL", "id": 72096, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/LEVACL/JuMP_j0XO80q.png", "date": "2025-10-03T11:00:00+02:00", "start": "11:00", "duration": "01:00", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-72096-optimization-with-jump", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/LEVACL/", "title": "Optimization with JuMP", "subtitle": "", "track": "Optimization", "type": "Tutorial", "language": "en", "abstract": "In this workshop, we'll focus on mathematical optimization using JuMP. You'll learn to write programs and solve integer linear, nonlinear, conic, and constraint optimization problems. We'll also discuss the recent new features of JuMP. Additionally, we'll detail important performance tips for solving large-scale models with JuMP.", "description": "In this workshop, we start with the basics of JuMP. We detail its syntax and design to help those new to the tool to get on board. We review the different classes of optimization problems that can be solved in JuMP and highlight its recent new features.\r\nWe then delve into a specific example: modeling an optimal power flow problem. For this particular case, we show how to model both a linear programming formulation and a nonlinear programming formulation with JuMP. Throughout the tutorial, we discuss important factors to ensure optimal performance, highlighting their importance in order for the program to scale for a large power flow network.", "recording_license": "", "do_not_record": false, "persons": [{"code": "MLJG3F", "name": "Beno\u00eet Legat", "avatar": "https://pretalx.com/media/avatars/MLJG3F_ntaDU6O.webp", "biography": "Beno\u00eet Legat is an Assistant Professor at UCLouvain in the Mathematical Engineering department (INMA) of the ICTEAM. He is working in optimization and is a core contributor of JuMP.", "public_name": "Beno\u00eet Legat", "guid": "bcf21464-fa9a-5dc0-80c4-03fbc9506174", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/MLJG3F/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LEVACL/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/LEVACL/", "attachments": []}, {"guid": "b3938c7d-6e6d-5a15-a28c-5aa57002b8cc", "code": "KMSGGQ", "id": 72106, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/KMSGGQ/WhatsApp__AqeWoIU.jpeg", "date": "2025-10-03T12:00:00+02:00", "start": "12:00", "duration": "00:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-72106-tulipaenergymodel-jl-tooling-for-the-energy-transition", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/KMSGGQ/", "title": "TulipaEnergyModel.jl - Tooling for the energy transition", "subtitle": "", "track": "Applications", "type": "Talk", "language": "en", "abstract": "Tulipa is a group of open-source tools for modelling and comparing European-scale energy systems, to support stakeholder decisions in the transition to sustainable energy.\r\n\r\nOne of the key challenges in energy system optimization modelling is achieving both high spatial-temporal detail and broad geographic or sectoral scope without overwhelming computational resources. Existing models are forced to prioritize either granularity or scale, because they cannot do both.\r\n\r\nTulipa addresses this trade-off through innovative modelling techniques that solve large, complex problems both *faster* and with *higher* accuracy \u2013 opening the door to analyses that were previously intractable.", "description": "In this talk, we will introduce **TulipaEnergyModel.jl**, discuss the core challenges in energy system optimization modelling, and explain core innovations and implementation choices in Tulipa to overcome these challenges.", "recording_license": "", "do_not_record": false, "persons": [{"code": "FNCEB3", "name": "Lauren Clisby", "avatar": "https://pretalx.com/media/avatars/FNCEB3_sxd0SfU.webp", "biography": "Lauren is a Program Manager at TNO, where they lead the development of open-source energy system modelling tools to support Europe's energy transition.", "public_name": "Lauren Clisby", "guid": "f815af88-0117-5956-a6b7-62af44453390", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/FNCEB3/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KMSGGQ/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KMSGGQ/", "attachments": []}, {"guid": "5dd36b3e-68ff-594c-a573-aa5947b1d1a9", "code": "Y8WPLD", "id": 70961, "logo": null, "date": "2025-10-03T12:30:00+02:00", "start": "12:30", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-70961-beyond-research-julia-as-the-core-of-commercial-space-mission-software", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y8WPLD/", "title": "Beyond Research: Julia as the Core of Commercial Space Mission Software", "subtitle": "", "track": "Applications", "type": "Lightning talk", "language": "en", "abstract": "This talk presents a research project turned product: a space mission simulator suite with Julia at its core. I'll highlight how the Julia language and ecosystem were essential to this effort, and how they offer a significant advantage over alternatives. I'll share an experience report about using Julia in production. I'll argue that Julia's potential for scientific tooling can be an edge in commercializing research \u2014 with proper attention to interoperability, tooling, and streamlined user experience.", "description": "Julia has earned recognition in scientific computing for its performance and expressiveness, with commercial adoption gaining momentum. This talk chronicles our experience transforming a space mission simulator stack from research prototype to  product.\r\n\r\nI'll introduce the problems we're trying to solve and our architectural approach, then demonstrate why Julia proved essential to our success. The presentation will highlight Julia's distinct advantages, including: rapid prototyping capabilities, accessibility for research engineers, interfaces for domain experts, performance, parallel computing support, language interoperability, and a powerful composable ecosystem for scientific computing.\r\n\r\nThe talk addresses practical aspects of using Julia in production and remaining challenges. I'll conclude by making the case that Julia offers significant potential for companies developing science-based products.", "recording_license": "", "do_not_record": false, "persons": [{"code": "U8ZGG9", "name": "Orestis Ousoultzoglou", "avatar": "https://pretalx.com/media/avatars/U8ZGG9_vTM7Uiy.webp", "biography": "space stuff", "public_name": "Orestis Ousoultzoglou", "guid": "3f83ed7a-c9a3-5623-a2e8-c74a79828a70", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/U8ZGG9/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y8WPLD/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y8WPLD/", "attachments": []}, {"guid": "ffb2a3a8-cb97-5058-885b-ae13e87d1094", "code": "BC9GHN", "id": 79780, "logo": null, "date": "2025-10-03T12:40:00+02:00", "start": "12:40", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79780-eutelsat-using-julia-to-simulate-the-present-and-the-future-of-satellite-communications", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/BC9GHN/", "title": "Eutelsat - Using Julia to simulate the present and the future of satellite communications", "subtitle": "", "track": "General", "type": "Sponsor talk", "language": "en", "abstract": "Sponsor talk by Eutelsat (https://www.eutelsat.com/).", "description": "-", "recording_license": "", "do_not_record": false, "persons": [{"code": "G3DWCG", "name": "Frank Augrandjean", "avatar": null, "biography": null, "public_name": "Frank Augrandjean", "guid": "47176dd3-2831-5d51-9b0c-c39fb8dab04c", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/G3DWCG/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/BC9GHN/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/BC9GHN/", "attachments": []}, {"guid": "7f2d7a60-ce72-5be1-bdf8-fc0cf9566704", "code": "T7YFDF", "id": 79887, "logo": null, "date": "2025-10-03T12:50:00+02:00", "start": "12:50", "duration": "00:10", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-79887-closing", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/T7YFDF/", "title": "Closing", "subtitle": "", "track": "General", "type": "Lightning talk", "language": "en", "abstract": "-", "description": "-", "recording_license": "", "do_not_record": false, "persons": [{"code": "PJVBLC", "name": "Guillaume Dalle", "avatar": "https://pretalx.com/media/avatars/PJVBLC_t336z2c.webp", "biography": null, "public_name": "Guillaume Dalle", "guid": "2515531a-5214-559d-968e-e8e0d2adc883", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PJVBLC/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/T7YFDF/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/T7YFDF/", "attachments": []}, {"guid": "894025bb-f2bf-5d94-b9bf-27116f585e49", "code": "JKQ88E", "id": 80029, "logo": null, "date": "2025-10-03T14:30:00+02:00", "start": "14:30", "duration": "03:30", "room": "Jean-Baptiste Say Amphitheater", "slug": "juliacon-local-paris-2025-80029-hackathon", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/JKQ88E/", "title": "Hackathon", "subtitle": "", "track": "General", "type": "Hackathon", "language": "en", "abstract": "Hacking together on our favorite Julia projects!", "description": "-", "recording_license": "", "do_not_record": false, "persons": [], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/JKQ88E/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/JKQ88E/", "attachments": []}], "Robert Faure Amphitheater": [{"guid": "ec1c6f64-d72f-5ddf-95d0-78190d8e543f", "code": "NUFEKB", "id": 67223, "logo": null, "date": "2025-10-03T09:30:00+02:00", "start": "09:30", "duration": "00:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-67223-property-testing-julia-code-or-how-to-overcome-bugs-and-learn-to-love-randomness", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/NUFEKB/", "title": "Property Testing Julia Code - or: How to overcome bugs and learn to love randomness", "subtitle": "", "track": "Language & teaching", "type": "Talk", "language": "en", "abstract": "Software testing is a complicated endeavor, which even with good code coverage often leaves doubts about the quality of the written code when subsequent usage still uncovers bugs. In this talk, we are going to take a look at some problems of traditional testing, challenges when it comes to testing Julia code in particular and how Property Based Testing with Supposition.jl can help with that.", "description": "Any experienced developer knows that 100% code coverage doesn't mean code is 0% buggy. A more comprehensive testing method called Property Based Testing (or PBT for short) can help bridge the gap between test cases and reality. Instead of only looking at one input per test with test driven design (TDD), PBT is geared towards looking at ALL possible inputs. \r\n\r\nWe're going to learn about PBT, how this approach to testing can help uncover hidden assumptions and uncover surprising behavior of well-known & frequently used functions from Base. For this, we'll use functionality from Supposition.jl, showcasing some of its capabilities in the process.", "recording_license": "", "do_not_record": false, "persons": [{"code": "9LYSWV", "name": "Valentin Bogad", "avatar": null, "biography": "Valentin Bogad is a Software Engineer from Vienna, who has been writing Julia code for close to 9 years. With a keen interest in Programming Language Semantics and too many buggy programs in various languages (among which, Java, C#, C and C++) to be embarrassed about writing buggy code, he is always on the hunt for better techniques to write high-quality software.", "public_name": "Valentin Bogad", "guid": "ec31ede9-9bed-5fdf-bf95-33f9af56edb2", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/9LYSWV/"}], "links": [{"title": "Documentation of Supposition.jl", "url": "https://seelengrab.github.io/Supposition.jl/stable/", "type": "related"}], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NUFEKB/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NUFEKB/", "attachments": []}, {"guid": "c268e26a-3eb5-516e-929a-12421ed378ec", "code": "Y3VWAA", "id": 70580, "logo": null, "date": "2025-10-03T10:00:00+02:00", "start": "10:00", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70580-argus-jl-matching-and-transforming-julia-syntax-and-writing-linting-rules", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y3VWAA/", "title": "Argus.jl: Matching and transforming Julia syntax and writing linting rules", "subtitle": "", "track": "Language & teaching", "type": "Lightning talk", "language": "en", "abstract": "This presentation introduces [Argus.jl](https://github.com/iuliadmtru/Argus.jl), a `JuliaSyntax`-based package that proposes a new approach to syntax manipulation in Julia. It draws inspiration from the Racket libraries [`syntax/parse`](https://docs.racket-lang.org/syntax/stxparse.html) and [`resyntax`](https://docs.racket-lang.org/resyntax/index.html).", "description": "Argus implements a framework for writing static analysis rules on top of a syntax matching mechanism. It is structured around several core concepts:\r\n\r\n  - Syntax patterns\r\n  - Pattern variables\r\n  - Syntax classes\r\n  - Syntax templates\r\n  - Rules\r\n\r\n**Syntax patterns** form the basis for syntax matching and closely resemble Julia code. For example, `@pattern x = 2` matches an assignment where the left-hand side is a variable named `x` and the right-hand side is the literal `2`. On the other hand, `@pattern {x} = 2` matches any assignment or short-form function definition where the right-hand side is the literal `2`; the expression on the left-hand side is bound to the _pattern variable_ `x`.\r\n\r\nA **pattern variable** is one of several special forms permitted within patterns. It can be seen as a \"hole\" that is filled by matching syntax. For example, when matching `@pattern {x} = 2` against the expression `f(a) = 2`, `x` is bound to `f(a)`. The result of a pattern match is either a set of bindings corresponding to the syntax matched by each pattern variable, or an error explaining why the matching failed.\r\n\r\nA pattern variable can be constrained by a **syntax class**. In the example above, `@pattern {x} = 2` is equivalent to `@pattern {x:::expr} = 2`, where `expr` is the syntax class that matches any expression. Syntax classes are defined through patterns and can reference other syntax classes. For example, a syntax class matching any assignment may be defined as such:\r\n\r\n```julia\r\nassign = @syntax_class \"assignment\" begin\r\n    @pattern {lhs:::identifier} = {rhs}\r\nend\r\n```\r\n\r\nArgus provides a set of pre-defined syntax classes, including `expr`, `identifier` and `assign`.\r\n\r\n**Syntax templates** are expanded to produce Julia code. They contain variables that are replaced with information gathered during pattern matching.\r\n\r\nA **rule** contains a description and a pattern. In the case of rules, matching recursively traverses a given unit of source code (e.g. a file) and collects the sub-expressions that match the rule's pattern. Pattern variables bound by these identifications are returned in corresponding _binding sets_.\r\n\r\nRules may be organised into _rule groups_. For example, it may be useful to group all rules related to Julia usage in a `lang` group:\r\n\r\n```julia\r\nlang_rules = RuleGroup(\"lang\")\r\n\r\n@define_rule_in_group lang_rules \"compare-nothing\" begin\r\n    description = \"\"\"\r\n    Comparisons to `nothing` should use `===`, `!==` or `isnothing`.\r\n    \"\"\"\r\n\r\n    pattern = @pattern begin\r\n        ~or(\r\n            nothing == {_},\r\n            {_} == nothing,\r\n            nothing != {_},\r\n            {_} != nothing\r\n        )\r\n    end\r\nend\r\n```\r\n\r\nThis presentation is an opportunity to share Argus with the community and exchange feedback and ideas for further improvement.\r\n\r\nGitHub: https://github.com/iuliadmtru/Argus.jl", "recording_license": "", "do_not_record": false, "persons": [{"code": "CUR3RR", "name": "Iulia Dumitru", "avatar": null, "biography": null, "public_name": "Iulia Dumitru", "guid": "2d43057c-6d2a-5bc7-ac43-bf744a4768df", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/CUR3RR/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y3VWAA/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Y3VWAA/", "attachments": []}, {"guid": "7485c723-a639-5fa2-97a8-c9038b608d1c", "code": "A3TYBX", "id": 72160, "logo": null, "date": "2025-10-03T10:10:00+02:00", "start": "10:10", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-72160-fixedsizearrays-jl-what-array-probably-should-have-been", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/A3TYBX/", "title": "FixedSizeArrays.jl: What Array probably should have been", "subtitle": "", "track": "Language & teaching", "type": "Lightning talk", "language": "en", "abstract": "In this talk we will introduce `FixedSizeArrays.jl`, a new package which implements mutable fixed-size arrays, based on the recent `Memory` type.", "description": "Julia's built-in `Array` type is a convenient multi-dimensional container for continuous data, with the ability to resize it. But do you really need this flexibility all the time? In most linear algebra-related applications you don't need to change the size of existing vectors or matrices, and the resizability comes at the cost of preventing potential compiler optimizations based on the known shape of arrays.\r\n\r\n`FixedSizeArray`s look very similar to Julia's standard `Array`s, with the difference that their size can't be changed, but unlike `StaticArrays.jl` the shape of the array isn't part of their type.\r\n\r\nWe will explore the opportunities for optimization and improved performance that `FixedSizeArrays.jl` offers, and outline possible future direction, including the possibility of having such an implementation in Julia `Base`.", "recording_license": "", "do_not_record": false, "persons": [{"code": "ST7KZT", "name": "Mos\u00e8 Giordano", "avatar": "https://pretalx.com/media/avatars/ST7KZT_3IdkVo1.webp", "biography": "Research Software Developer at UCL during the day, binary builder during the night.", "public_name": "Mos\u00e8 Giordano", "guid": "0bdcf3dd-f493-5992-a04c-54716e0accef", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/ST7KZT/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/A3TYBX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/A3TYBX/", "attachments": []}, {"guid": "72f988f4-ac9b-5b01-b194-b03ccf0ffe9b", "code": "VCLPPT", "id": 70953, "logo": null, "date": "2025-10-03T10:20:00+02:00", "start": "10:20", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70953-juliacon-proceedings-behind-the-scenes", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/VCLPPT/", "title": "JuliaCon Proceedings: behind the scenes", "subtitle": "", "track": "Language & teaching", "type": "Lightning talk", "language": "en", "abstract": "JuliaCon has its own peer-reviewed academic proceedings journal, where authors can publish their work presented at JuliaCon. This talk will give an overview of the JuliaCon proceedings, discussing its structure, its submission and review process, its technical infrastructure and how people in the community can get involved.", "description": "JuliaCon often features researchers presenting new tools or results valuable to scientific communities. To support academic recognition through publications, JuliaCon launched its peer-reviewed journal, JuliaCon Proceedings, in 2019.\r\n\r\nInspired by the Journal of Open Source Software (JOSS), JuliaCon Proceedings is completely free and follows a transparent, open peer-review process. Reviews take place on GitHub as issues, where reviewers and authors openly exchange feedback. The journal accepts both extended abstracts and full papers, and all JuliaCon presenters are eligible to submit their talks for publication.\r\n\r\nThis talk will outline the journal\u2019s structure, submission and review process, underlying technical infrastructure, and opportunities for community involvement.", "recording_license": "", "do_not_record": false, "persons": [{"code": "ZTATFJ", "name": "Ludovic R\u00e4ss", "avatar": "https://pretalx.com/media/avatars/ZTATFJ_PLThLkL.webp", "biography": "Computational geoscientists with Earth science background. Julia GPU and HPC enthusiast.", "public_name": "Ludovic R\u00e4ss", "guid": "848e7813-344c-5834-aeeb-a0d80906a475", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/ZTATFJ/"}, {"code": "EUWBHH", "name": "Luca Ferranti", "avatar": null, "biography": "hi", "public_name": "Luca Ferranti", "guid": "67f503a8-48f7-5b7f-b7b2-8813293a497a", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/EUWBHH/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VCLPPT/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VCLPPT/", "attachments": []}, {"guid": "d41ae3eb-a559-5c80-913d-139e41034364", "code": "M3C97F", "id": 72270, "logo": null, "date": "2025-10-03T11:00:00+02:00", "start": "11:00", "duration": "00:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-72270-what-s-new-with-bifurcationkit", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/M3C97F/", "title": "What's new with BifurcationKit", "subtitle": "", "track": "Error, derivatives, stability", "type": "Talk", "language": "en", "abstract": "BifurcationKit.jl is a powerful and flexible Julia package for the numerical study of bifurcations in nonlinear systems. This tutorial will guide participants through the practical use of BifurcationKit.jl to analyze dynamical systems, including fixed points, bifurcation diagrams, continuation of solutions, and codimension-1 bifurcations. Through interactive coding examples, participants will learn how to apply the package to both low- and high-dimensional systems and understand the computational tools behind continuation and stability analysis.", "description": "The aim of this tutorial is to introduce the core concepts and workflows of BifurcationKit.jl through hands-on examples, targeting researchers, students, and engineers interested in nonlinear dynamics, pattern formation, and bifurcation theory.\r\n\r\nIntended Audience\r\n\t\r\n-\tApplied mathematicians and physicists working with dynamical systems\r\n-\tJulia users interested in modeling and simulation\r\n-\tGraduate students learning bifurcation theory or numerical continuation\r\n-\tUsers of packages like DifferentialEquations.jl or DynamicalSystems.jl\r\n\r\nPrerequisites\r\n\r\n- Basic knowledge of Julia (functions, packages, plotting)\r\n- Familiarity with differential equations and dynamical systems is helpful but not required\r\n\r\nOutline\r\n\r\n1. Introduction (5 minutes)\r\n\r\n- Motivation for bifurcation analysis\r\n- Overview of BifurcationKit.jl capabilities and its ecosystem\r\n\r\n2. Getting Started (5 minutes)\r\n\r\n- Installing and importing BifurcationKit\r\n- Defining a simple bifurcation problem\r\n\r\n3. Continuation Methods (10 minutes)\r\n\r\n- Pseudo-arclength continuation and others\r\n- Tracking solution branches\r\n- Plotting bifurcation diagrams\r\n\r\n4. Detecting and Analyzing Bifurcations (15 minutes)\r\n\r\n- Stability and eigenvalue tracking\r\n- Saddle-node, Hopf, branch point, etc\r\n- Codimension-1 bifurcation examples\r\n- branching from bifurcations (branch / Hopf points)\r\n- Integration with DifferentialEquations.jl and modeling tools\r\n\r\n5. Focus on periodic orbits (10 minutes)\r\n\r\n- Methods for computing them\r\n- Branching from bifurcations\r\n- Codimension-1 bifurcation examples\r\n\r\n6. Applications and Advanced Topics (15 minutes)\r\n\r\n- PDE continuation with spectral methods\r\n- Using Newton-Krylov solvers for large systems\r\n\r\n7. Q&A and Discussion (5 minutes)\r\n\r\n- Troubleshooting tips\r\n- Questions from participants\r\n- Ideas for extensions and open problems", "recording_license": "", "do_not_record": false, "persons": [{"code": "PGWM93", "name": "Romain VELTZ", "avatar": null, "biography": null, "public_name": "Romain VELTZ", "guid": "5c3c82c4-fe04-5808-a9c7-3eb1d2476334", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PGWM93/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/M3C97F/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/M3C97F/", "attachments": []}, {"guid": "40b3fc5d-a8b6-5568-92c4-18fd18a3c818", "code": "KTGPQE", "id": 70971, "logo": null, "date": "2025-10-03T11:30:00+02:00", "start": "11:30", "duration": "00:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70971-leveraging-sparsity-to-accelerate-automatic-differentiation", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/KTGPQE/", "title": "Leveraging Sparsity to Accelerate Automatic Differentiation", "subtitle": "", "track": "Error, derivatives, stability", "type": "Talk", "language": "en", "abstract": "Jacobians and Hessians play vital roles in scientific computing and machine learning, from optimization to probabilistic modeling. While these matrices are often considered too computationally expensive to calculate, their inherent sparsity can be leveraged to dramatically accelerate Automatic Differentiation (AD). By building on top of DifferentiationInterface.jl, we are able to bring Automatic Sparse Differentiation to all major Julia AD backends, including ForwardDiff and Enzyme.", "description": "Despite AD's widespread adoption in the Julia ecosystem, Automatic Sparse Differentiation (ASD) remains an underutilized technique. This talk explores the key components of ASD, beginning with a concise overview of the theory behind conventional AD. We then dive into the two main components of ASD: sparsity pattern detection and matrix coloring.\r\n\r\nWe then introduce a novel Julia ASD pipeline built on three open-source packages:\r\n1. [SparseConnectivityTracer.jl](https://github.com/adrhill/SparseConnectivityTracer.jl) - implements sparsity pattern detection via operator-overloading  \r\n2. [SparseMatrixColorings.jl](https://github.com/gdalle/SparseMatrixColorings.jl) - implements matrix coloring algorithms\r\n3. [DifferentiationInterface.jl](https://github.com/JuliaDiff/DifferentiationInterface.jl) - provides a unified interface for AD\r\n\r\nThrough DifferentiationInterface.jl, we bring ASD capabilities to all major Julia AD backends, including ForwardDiff.jl, ReverseDiff.jl, Zygote.jl, and Enzyme. We hope this work can serve as a blueprint for future backend agnostic AD improvements.\r\nThe presentation concludes with practical demonstrations, featuring detailed performance benchmarks and clear guidelines for choosing between ASD and traditional AD approaches.", "recording_license": "", "do_not_record": false, "persons": [{"code": "PJVBLC", "name": "Guillaume Dalle", "avatar": "https://pretalx.com/media/avatars/PJVBLC_t336z2c.webp", "biography": null, "public_name": "Guillaume Dalle", "guid": "2515531a-5214-559d-968e-e8e0d2adc883", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/PJVBLC/"}, {"code": "3CFNUV", "name": "Adrian Hill", "avatar": "https://pretalx.com/media/avatars/3CFNUV_0BGgi61.webp", "biography": "PhD student in the Machine Learning Group at TU Berlin.\r\nInterested in in automatic differentiation, explainability and dynamical systems.\r\n\r\n- Personal website: [adrianhill.de](https://adrianhill.de)\r\n- GitHub: [@adrhill](https://github.com/adrhill)\r\n- [Project spotlight](https://adrianhill.de/projects/)", "public_name": "Adrian Hill", "guid": "62c9975b-2116-5458-a96c-e00cb8332090", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/3CFNUV/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KTGPQE/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KTGPQE/", "attachments": []}, {"guid": "96523116-3dac-513c-9bc8-919e6ed9cead", "code": "8QB9EN", "id": 67043, "logo": null, "date": "2025-10-03T12:00:00+02:00", "start": "12:00", "duration": "00:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-67043-algorithmic-differentiation-with-mooncake-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/8QB9EN/", "title": "Algorithmic Differentiation with Mooncake.jl", "subtitle": "", "track": "Error, derivatives, stability", "type": "Talk", "language": "en", "abstract": "[Mooncake.jl](https://github.com/compintell/Mooncake.jl/) is an algorithmic differentiation (AD) tool written in Julia. It is characterised by support for a wider range of Julia language features than existing tools, and performance which is typically better than comparable tools written in Julia. It has extensive documentation, simple-to-use tools for correctness testing supported by a precise type system, and is best used via [DifferentiationInterface.jl](https://github.com/gdalle/DifferentiationInterface.jl). In this talk I will attempt to both justify and qualify these claims, and will conclude with an opinionated outlook on the future of AD in Julia.", "description": "# Overview\r\n\r\nIn order to justify and qualify the claims made in the abstract, this talk will explain the functionality that [Mooncake.jl](https://github.com/compintell/Mooncake.jl/) offers to users, and provide a broad-brush explanation of how Mooncake.jl works -- for example what Julia objects it transforms, what it does to them, and what it returns. Below I discuss some more detail around particular points that I shall address.\r\n\r\n## Language Feature Support\r\n\r\nIn order to develop the claim regarding Mooncake.jl's breadth of support for the Julia language, I will explain which Julia language features are supported by Mooncake.jl in comparison with [Zygote.jl](https://github.com/FluxML/Zygote.jl), [ReverseDiff.jl](https://github.com/JuliaDiff/ReverseDiff.jl), and [Enzyme.jl](https://github.com/EnzymeAD/Enzyme.jl).\r\n\r\n## Performance\r\n\r\nComparing the performance of different AD systems is a tricky business -- one will always be able to find counterexamples to any claim that one system is faster than another. Therefore, I'll just try to offer some performance comparisons on specific examples, highlighting notable performance implications of the way Mooncake.jl goes about implementing AD. The aim is to build intuition for the kinds of situations in which Mooncake.jl is faster or slower than other systems, and by roughly how much.\r\n\r\n## Rule System and Robustness\r\n\r\nMooncake.jl has a rule system which appears quite similar to that of [ChainRules.jl](https://github.com/JuliaDiff/ChainRules.jl), but which differs in a couple of crucial ways. From a robustness perspective, the most notable difference is that any tangents / gradients associated to a value of type `P` must be of type `Mooncake.tangent_type(P)`. That is, each \"primal type\" is associated to a unique \"tangent type\". Conversely, ChainRules.jl takes a laissez-faire approach, in which no such restrictions are imposed. I will explain this in more depth, and its implications for users, library authors who wish to ensure that Mooncake.jl can differentiate their code, and maintainers of Mooncake.jl itself.\r\n\r\nDespite these differences, it is possible to re-use some rules from ChainRules.jl inside Mooncake.jl. I will briefly highlight how and where this is done.\r\n\r\n## Testing Tooling\r\n\r\nAD tools require thorough testing to have any hope of being correct. Since correctness errors are really the worst kind of errors that an AD tool can make, thorough testing is central to any AD tool. Consequently, testability has been at the core of all design decisions made while developing Mooncake.jl. I will explain how the tools we have implemented for testing are made possible by our design decisions, and how users and library authors can utilise them in their own code.", "recording_license": "", "do_not_record": false, "persons": [{"code": "LQKANJ", "name": "Will Tebbutt", "avatar": "https://pretalx.com/media/avatars/LQKANJ_frbJWzY.webp", "biography": "I am a postdoc at the Alan Turing Institute in London, having previously been a postdoc and PhD student in the Machine Learning Group in Cambridge. I am interested in probabilistic programming, Gaussian processes, algorithmic differentiation, and machine learning for weather forecasting.\r\n\r\nI have been a Julia user for a while. I have worked on the algorithmic differentiation ecosystem (Zygote.jl, ChainRules.jl, and Mooncake.jl), the various packages in the JuliaGaussianProcesses organisation. I have also been working closely with the Turing.jl team.", "public_name": "Will Tebbutt", "guid": "cc3690c7-eb98-5539-986a-3ac1ccba68de", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/LQKANJ/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/8QB9EN/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/8QB9EN/", "attachments": []}, {"guid": "d87be135-0eea-5af4-acd8-77fbbc75b137", "code": "NJLJTY", "id": 70905, "logo": "https://pretalx.com/media/juliacon-local-paris-2025/submissions/NJLJTY/03-ocean-r_yxXqWzO.png", "date": "2025-10-03T12:30:00+02:00", "start": "12:30", "duration": "00:10", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-70905-computational-storytelling-reimagining-scientific-communication-through-pluto-s-interactive-julia-notebooks", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/NJLJTY/", "title": "Computational Storytelling: Reimagining Scientific Communication through Pluto's Interactive Julia Notebooks", "subtitle": "", "track": "Language & teaching", "type": "Lightning talk", "language": "en", "abstract": "This project is a playful collection of interactive Pluto.jl notebooks, made to turn tricky scientific and math ideas into something explorable! The aim? To make learning feel less like a boring lecture and more a \u201cwow, that\u2019s cool!\u201d moment\u2014while sharing clean, high-quality Julia code anyone can use. I also want to take you behind the scenes to show how you can build these notebooks yourself using the same tools for computational storytelling *(and have fun doing it!)*.", "description": "Let\u2019s face it\u2014science can be kind of a snob. It hides behind jargon, paywalls, and tools so complicated they seem like you need a PhD just to hit \u201crun.\u201d Meanwhile, people today are hungry for knowledge\u2014curious, eager, and constantly Googling everything (as long as it doesn\u2019t require an attention span longer than a TikTok video). So why not meet them where they are? **That\u2019s where computational storytelling comes in.**\r\n\r\nWith [Pluto.jl](https://plutojl.org/) as our trusty sidekick and the straight-up simplicity of Julia, we\u2019re breaking down barriers with beginner-friendly, interactive notebooks that make big ideas in science and math feel fun, clear, and totally doable. Why Pluto? It blends live interactivity with real code\u2014so you get the best of both worlds. Everything updates in real time, making learning feel more like play. And that\u2019s exactly our mission! Together with Fons *(the brain behind the awesome Pluto project)*, I want to make learning joyful, code reusable, and science feel like something everyone can jump into right away. \r\n\r\n*To show what that looks like, I\u2019ve created a series of notebooks, each zooming in on one cool concept.*\r\n\r\nOne notebook dives into **domain coloring**, turning complex numbers into vibrant visual art. Math doesn\u2019t have to feel dry or intimidating\u2014so we brought out the colors to prove it. Curious? [Check it out!](https://featured.plutojl.org/math/domain%20coloring.html)\r\n\r\nAnother notebook unpacks **decision trees** in AI using a hands-on approach: tweak some sliders, change the inputs, and see how the model responds in real time. With so many people getting interested in AI these days, we wanted to offer a easy, friendly peek behind the curtain. [Try it yourself!](https://featured.plutojl.org/machine-learning/decision-trees.html)\r\n\r\nAnd then there\u2019s the story-driven **ozone repair** notebook, which uses interactive maps and animations to show how the world came together to solve a global crisis. It\u2019s a scientific win that deserves more love\u2014and while this one\u2019s still in progress, we\u2019ve dropped a sneak peek screenshot to give you a taste.\r\n\r\nEvery notebook is designed to feel less like a textbook and more like a little learning adventure, wrapped in approachable code and powered by curiosity. And we don\u2019t stop there\u2014we\u2019re also pulling back the curtain on how you can make such notebooks. \r\n\r\n*How can you use computation to tell a story, beyond just generating plots and figures? Can Julia code be part of your story?*\r\n\r\nIn the talk, I\u2019ll walk you through a few of these hand-picked notebooks, and we'll interactively learn about these topics together! Then I'll go into the topic of *Computational Storytelling* and give you a behind-the-scenes tour of what goes on in my head while building the notebooks, sharing tips, tricks, and the must-know do\u2019s and don\u2019ts of telling a great science story with code that apply across fields, to help you tell your next story the same way. \u200bStill curious? Let\u2019s make the world of science communication different together, let\u2019s make this talk happen!", "recording_license": "", "do_not_record": false, "persons": [{"code": "VAFRRY", "name": "Boshra Ariguib", "avatar": "https://pretalx.com/media/avatars/VAFRRY_lf0lX6G.webp", "biography": "Hi \u2600\ufe0f\u00a0I am a CS Masters student in my final semester at the University of Stuttgart and a big enthusiast of the Pluto.jl Project \ud83c\udf89\r\n\r\nI\u2019m passionate about making coding and science accessible to everyone, regardless of age or background. If you\u2019re working on something along those lines, I\u2019d love to chat! This is my very first conference, so I\u2019m especially excited to meet new people and make friends \ud83d\ude0a", "public_name": "Boshra Ariguib", "guid": "828efe95-6a73-5bd3-ad58-6a17b5f96b75", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/VAFRRY/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NJLJTY/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/NJLJTY/", "attachments": []}, {"guid": "266cd0c3-d5e8-532a-a977-b86064da2b75", "code": "Z333LG", "id": 80028, "logo": null, "date": "2025-10-03T14:30:00+02:00", "start": "14:30", "duration": "03:30", "room": "Robert Faure Amphitheater", "slug": "juliacon-local-paris-2025-80028-hackathon", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/Z333LG/", "title": "Hackathon", "subtitle": "", "track": "General", "type": "Hackathon", "language": "en", "abstract": "Hacking together on our favorite Julia projects!", "description": "-", "recording_license": "", "do_not_record": false, "persons": [], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Z333LG/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/Z333LG/", "attachments": []}], "Coffee room": [{"guid": "3f3cc587-1711-5b4a-9147-baa3fe114936", "code": "VUXCUM", "id": 69853, "logo": null, "date": "2025-10-03T10:30:00+02:00", "start": "10:30", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-69853-carnotcycles-jl-a-package-to-simulate-thermodynamic-cycles", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/VUXCUM/", "title": "CarnotCycles.jl -  A package to simulate thermodynamic cycles", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "CarnotCycles.jl provides a base framework to model thermodynamic cycles in Julia. The name is inspired by Nicolas L\u00e9onard Sadi Carnot who among many other things theorized the most efficient power cycle called the Carnot\u2019s cycle. This package aims to have the versatility of modeling various thermodynamic cycles, ranging from Vapour compression cycles to Organic Rankine cycle.", "description": "This package allows to model a thermodynamic cycle followed by the relevant phase plots. It is based on acausal framework where users can make their own components and check the relevant performance of the system. As of now it is made for steady state calculations.", "recording_license": "", "do_not_record": false, "persons": [{"code": "QEBWLM", "name": "Sushrut Deshpande", "avatar": null, "biography": "I am a student at the Royal Military Academy, Belgium. \r\nI am interested in thermodynamics and numerical techniques", "public_name": "Sushrut Deshpande", "guid": "4c656416-7e9b-51fd-959c-96f928650aa1", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/QEBWLM/"}], "links": [{"title": "Repository", "url": "https://github.com/Sush1090/CarnotCycles.jl", "type": "related"}], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VUXCUM/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/VUXCUM/", "attachments": []}, {"guid": "441a46ae-de10-50b8-a4bd-0c333477140c", "code": "ZEJU7T", "id": 70691, "logo": null, "date": "2025-10-03T10:33:00+02:00", "start": "10:33", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70691-numeric-integration-on-simplicies-and-orthotopes-using-hadaptiveintegration-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZEJU7T/", "title": "Numeric integration on simplicies and orthotopes using `HAdaptiveIntegration.jl`", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "We present the new package [`HAdaptiveIntegration.jl`](https://github.com/zmoitier/HAdaptiveIntegration.jl) which can compute the integral of functions over simplicies (triangle, tetrahedron, ...) and axis-aligned orthotopes (rectangle, cuboid, ...). The numerical computation of the integral is based on efficient tabulated embedded cubature rules and automatic mesh refinement (h-adaptivity) guided by error estimation given by the embedded cubature.", "description": "[`HAdaptiveIntegration.jl`](https://github.com/zmoitier/HAdaptiveIntegration.jl) is a Julia package designed for numerical integration over multidimensional domains. It computes integrals of the form *I* = \u222b *f*(*x*) d*x* over \u03a9 where *f* is any Julia function and \u03a9 represents domains such as simplices (triangle, tetrahedron, ...) and axis-aligned orthotopes (rectangle, cuboid, ...). The package employs an adaptive approach, dynamically refining the integration domain as needed. It uses embedded cubature rules to provide error estimates, aiming to achieve high accuracy while minimizing function evaluations.\r\n\r\nFeatures include:\r\n- Adaptive integration over simplices and axis-aligned orthotope of **any dimension**,\r\n- Utilization of **efficient tabulated cubatures** for low-dimensional simplices and orthotopes,\r\n- Support for custom embedded cubature rules,\r\n- Arbitrary precision arithmetic.", "recording_license": "", "do_not_record": false, "persons": [{"code": "MPXGDG", "name": "Moitier", "avatar": null, "biography": null, "public_name": "Moitier", "guid": "cd3f1976-928e-5525-913e-25fb77c7f0ef", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/MPXGDG/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZEJU7T/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ZEJU7T/", "attachments": []}, {"guid": "58f928c0-eb6b-5b13-8f93-254ccc6fb39c", "code": "KWJLG3", "id": 70670, "logo": null, "date": "2025-10-03T10:36:00+02:00", "start": "10:36", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70670-on-the-optimal-control-of-membrane-filtration-systems", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/KWJLG3/", "title": "On the optimal control of membrane filtration systems", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "The performance of membrane filtration system could be improved by optimal control. We present a code in Julia for the optimal control of membrane filtration systems, for a wide range of models proposed in the literature, mainly based on combination of automatic differentiation (ForwardDiff.jl) and numerical integration of ordinary differential equations (OrdinaryDiffEq.jl).", "description": "Membrane filtration systems are widely used in various industries, including water treatment, food processing, and pharmaceuticals. The performance of these systems is influenced by several factors, including membrane fouling, backwash cycle, and flow dynamics. Optimal control of these systems is crucial for maximizing efficiency and minimizing operational costs. Some recent works [1, 2] propose a model-based optimal control strategy. Our purpose is to develop a generic code to provide the optimal control strategy for a wide range of models of membrane filtration systems. The choice of the Julia programming language and its ecosystem is motivated by its high performance and ease of use for numerical integration, automatic differentiation, and optimization. The main features of the code will be presented, as well as the potential applications of this code in industry.\r\n\r\nReferences\r\n[1] Nesrine Kalboussi, Alain Rapaport, T\u00e9rence Bayen, Nihel Ben Amar, Fatma Ellouze, et al., Optimal control of membrane filtration systems, IEEE Transactions on Automatic Control, 2019.\r\n[2] Farouk Aichouche, Nesrine Kalboussi, Alain Rapaport, J\u00e9rome Harmand, Modeling and optimal control for production-regeneration systems - preliminary results, European Control Conference ECC\u201920, 2020.", "recording_license": "", "do_not_record": false, "persons": [{"code": "QAA8XV", "name": "Dutto", "avatar": "https://pretalx.com/media/avatars/QAA8XV_HLSUDdT.webp", "biography": "I am currently a research engineer at the INRAE Occitanie-Montpellier, working on the WOc-WoD project with J\u00e9r\u00f4me Harmand and Alain Rapaport. The main objective is to develop optimal control laws for membrane filtration systems. My fields of interests are optimal control, artificial intelligence and optimization.", "public_name": "Dutto", "guid": "616fe771-bfb3-53d2-b6fb-7c73803467d9", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/QAA8XV/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KWJLG3/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/KWJLG3/", "attachments": []}, {"guid": "6a89fb0c-7d2d-5ce9-8951-22fc1c45faa5", "code": "ANVBF7", "id": 72233, "logo": null, "date": "2025-10-03T10:39:00+02:00", "start": "10:39", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72233-practical-identifiability-and-predictability-analysis-in-julia", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/ANVBF7/", "title": "Practical Identifiability and Predictability Analysis in Julia", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Learn how to perform practical identifiability and predictability analysis in Julia. This talk showcases how LikelihoodProfiler.jl fits into the model calibration workflow, integrates with the SciML ecosystem, and leverages Julia's parallel computing features. A Systems Biology model will be used to illustrate the analysis and demonstrate the overall workflow.", "description": "This talk presents recent features of LikelihoodProfiler.jl, a Julia package for practical identifiability and uncertainty analysis in parameterized models. We walk through a modeling example from Systems Biology that highlights how the profile likelihood approach can assess both parameter identifiability and model predictability. We also demonstrate the variety of profiling methods, their compatibility with SciML interfaces, support for the PEtab format, and parallel execution. The goal is to show how LikelihoodProfiler.jl fits into model calibration workflows.", "recording_license": "", "do_not_record": false, "persons": [{"code": "DMYTV9", "name": "Ivan Borisov", "avatar": "https://pretalx.com/media/avatars/DMYTV9_d2zv9jO.webp", "biography": "Ivan Borisov is a software developer at InSysBio CY. His work focuses on the development and enhancement of mathematical and computational methods in Systems Biology and Quantitative Systems Pharmacology.", "public_name": "Ivan Borisov", "guid": "3c480b61-3eba-51a8-ac7f-0abfff417171", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/DMYTV9/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ANVBF7/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/ANVBF7/", "attachments": [{"title": "Poster", "url": "/media/juliacon-local-paris-2025/submissions/ANVBF7/resources/_4kOHMEd.pdf", "type": "related"}]}, {"guid": "f88451bc-069f-582b-9c39-060b4bf0aa9c", "code": "RXJEAJ", "id": 70594, "logo": null, "date": "2025-10-03T10:42:00+02:00", "start": "10:42", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70594-qpbamm-jl-a-parallelizable-admm-approach-for-block-structured-quadratic-programs", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/RXJEAJ/", "title": "qpBAMM.jl: a parallelizable ADMM approach for block-structured quadratic programs", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "In this talk we will present the ADMM-based quadratic program (QP) solver qpBAMM.jl. This solver is particularly suitable for solving blocked-structured QPs that occur in the context of the direct approach to solving optimal control problems. Besides the method used in the solver, we want to present the usage of our solver, as well as numerical results compared to other augmented Lagrangian based QP solvers available in Julia.", "description": "Block-structured quadratic programs (QPs) frequently arise in the context of the direct approach to solving optimal control problems. For successful application of direct optimal control algorithms to many real-world problems it is paramount that these QPs can be solved efficiently and reliably. Besides interior-point methods and active-set methods, ADMM-based quadratic programming approaches have gained popularity. Existing ADMM-based QP solvers typically split the problem into an equality-constrained QP and a projection problem onto, for example, a box-shaped feasible region. While this is a proven and widely used method, splitting the optimal control QP in this way does not easily allow to utilize and exploit the block structure of optimal control QPs to the fullest extent. We address this situation by splitting the QP into a box-constrained QP, and a projection onto the linearized dynamics. Restricting the QP constraint set to a box, the solution becomes possible through an embarrassingly parallel gradient projection method. The projection step can be parallelized only partially, but benefits from the invariance of the constraint set to be projected on. With the proposed splitting, we are hence able to utilize the optimal control block structure, and solve relevant parts of the problem in parallel. We offer an implementation in Julia, present numerical results on random problems as well as on a popular benchmark problem, and demonstrate significant advantages over recent Augmented Lagrangian method based solvers when applied to instances of optimal control problems with many state variables.", "recording_license": "", "do_not_record": false, "persons": [{"code": "P79BSC", "name": "Michel Lahmann", "avatar": null, "biography": "PhD student @ Technische Universit\u00e4t Braunschweig, Institute for Mathematical Optimization", "public_name": "Michel Lahmann", "guid": "9e964f3d-3cce-5f17-870f-3e98712798cc", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/P79BSC/"}, {"code": "DLXKPP", "name": "Martin K\u00f6hler", "avatar": null, "biography": "PhD student at TU Braunschweig, Germany, interested in large-scale network control, PDE constrained optimization and nonlinear solvers in Julia.", "public_name": "Martin K\u00f6hler", "guid": "1b35e515-30c4-5f7c-85b3-722e8f43bfae", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/DLXKPP/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RXJEAJ/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/RXJEAJ/", "attachments": []}, {"guid": "17c22990-0d7f-5881-b62d-6f92625299d1", "code": "WQUAZX", "id": 72257, "logo": null, "date": "2025-10-03T10:45:00+02:00", "start": "10:45", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-72257-three-julian-years-at-ienai-space", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/WQUAZX/", "title": "Three Julian years at IENAI SPACE", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "At IENAI SPACE, we adopted Julia 3 years ago to develop our mission analysis tools for simulation and optimization of small-satellite space missions. Thanks to high-level productivity, strong performance and a vibrant open-source ecosystem with strong foundations (highlighting DifferentialEquations.jl and SatelliteToolbox.jl in particular), we are quickly achieving our objectives with a small team. We will discuss strengths, pain points, and highlight some advanced analyses we perform with them", "description": "This talk aims to describe how Julia and its ecosystem has enabled IENAI SPACE to quickly develop an cutting-edge mission analysis toolbox with a small team. \r\n\r\nBackground\r\n----------\r\n\r\nWe are a NewSpace startup based on the Madrid metropolitan area that aims to address the mobility gap in small satellites through both hardware solutions (in the form of the Athena electrospray thruster) and software solutions (such as IENAI 360, the main subject of this talk) that allow satellite and constellation operators to efficiently perform analysis and design of small-satellite missions; we have a particular focus on low-thrust propulsion systems, whose operational implications often demand specific analysis solutions and methodologies that are not as easily covered by traditional tools.\r\n\r\nStructure\r\n---------\r\n\r\n1) Introduction: we will describe our mission (helping to solve the problem of small-satellite propulsion and the design of their mobility operations) and how a pillar of our solutions is the development of accurate, performant, versatile and effective mission analysis software.\r\n\r\n2) History and ecosystem foundations - in other words, why and when did we choose Julia? We explain how the assessed benefits, costs, and risks were a good fit for a startup such as us. In particular, even though the use of a newer programming language of smaller industrial popularity entails some risks and drawbacks, we considered that they were outweighed by the combination of high productivity, good performance, good versatility, and a very strong ecosystem for our specific context. In particular, we will discuss some critical enablers for us, such as the existence of the DifferentialEquations.jl suite (and the SciML ecosystem in general) and the SatelliteToolbox.jl library, which covers many pillars and models in the space analysis field.\r\n\r\n3) Solutions and analysis examples - at this stage, we will discuss the main features and architecture of our modular mission analysis solution, 360, whose core computational engine is written in Julia. We will direct particular emphasis to how multiple dispatch enables powerful abstraction patterns, which allow us to implement a versatile panoply of functionalities without sacrificing performance when it matters. We will discuss deterministic and uncertain simulation use cases, parametric analysis, and multi-objective optimization.\r\n\r\n4) Practical aspects of working with Julia: we will conclude with some commentary on the benefits, drawbacks and pain points we encountered and encounter in our Julian journey, focusing not just on the purely technical aspects but also on the socio-laboral implications of this choice from our experience.", "recording_license": "", "do_not_record": false, "persons": [{"code": "M3FYQD", "name": "Daniel Gonz\u00e1lez Arribas", "avatar": "https://pretalx.com/media/avatars/M3FYQD_hzx3y5w.webp", "biography": "Daniel Gonz\u00e1lez Arribas is a control & automation engineer specializing in computational modeling, simulation and optimization of aerospace system. He obtained his PhD on Aerospace Engineering in 2019 at Universidad Carlos III de Madrid, where he performed research on aircraft trajectory optimization and optimal flight planning under meteorological uncertainty. He is now at IENAI SPACE, a NewSpace startup, where he leads the development of the computational engine of IENAI's mission analysis tools.", "public_name": "Daniel Gonz\u00e1lez Arribas", "guid": "ce887233-7b04-5849-b8a4-c0b61fe66d25", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/M3FYQD/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/WQUAZX/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/WQUAZX/", "attachments": []}, {"guid": "55d995f1-d832-528c-8ec7-9cd6370e45be", "code": "GAD8AS", "id": 70784, "logo": null, "date": "2025-10-03T10:48:00+02:00", "start": "10:48", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-70784-turing-jl-what-s-new", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/GAD8AS/", "title": "Turing.jl: What's new?", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Turing.jl is a probabilistic programming language written entirely in Julia. It makes writing statistical models and performing Bayesian inference on them easy. Turing.jl is used in a wide range of fields from astrophysics to epidemiology, and integrates with the Julia scientific computing ecosystem. It has existed since Julia v0.4, and in this poster we provide an update on its latest new features and current plans for its development.", "description": "Turing.jl has grown organically over the 9 years of its existence, and now has an estimated user base of some thousands. We are actively developing it towards a stable v1.0 release with a small team of full-time software engineers and researchers and community contributions. More than a single package, Turing.jl is really an ecosystem of interoperating packages, including an abstract interface for Markov Chain Monte Carlo, many different sampling algorithms, a domain specific language for writing statistical models, and others. It also integrates well with many other packages in the Julia scientific computing ecosystem, which is one of its key strengths compared to other probabilistic programming languages.\r\n\r\nSince our last update at JuliaCon 2024, we have made significant changes to\r\n* Handling of submodels, i.e. statistical models as subparts of other statistical models.\r\n* Support for different automatic differentiation packages.\r\n* Several big changes to the interface for inference algorithm developers.\r\n* Better support for variational inference.\r\n\r\nOur poster summarises these key new developments, as well as any other significant changes we make between writing this description and presenting the poster. We also describe our roadmap for the future and invite everyone to come talk to us about their use of Turing.jl.", "recording_license": "", "do_not_record": false, "persons": [{"code": "HKWNAZ", "name": "Markus Hauru", "avatar": null, "biography": "I am a Senior Research Data Scientist at the Alan Turing Institute, working on Turing.jl development. In the past I have also contributed to the Julia ecosystem of tensor network packages.", "public_name": "Markus Hauru", "guid": "3c8f4871-b341-5394-8920-1b87d7cd46b8", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/HKWNAZ/"}, {"code": "F7NKSD", "name": "Penelope Yong", "avatar": null, "biography": "she/her\r\n\r\nEx-quantum chemist but nowadays more interested in Jane Austen, my Pokemon collection, classical music, functional programming, or waterfowl.\r\n\r\nSometimes I dabble with Turing.jl.", "public_name": "Penelope Yong", "guid": "64de1b1a-5dbb-5168-96cd-17014964d75d", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/F7NKSD/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/GAD8AS/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/GAD8AS/", "attachments": []}, {"guid": "826cee79-1395-5f14-99b5-522138c51224", "code": "DJCV9D", "id": 69574, "logo": null, "date": "2025-10-03T10:51:00+02:00", "start": "10:51", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-69574-type-stable-heterogeneous-arrays", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/DJCV9D/", "title": "Type-stable heterogeneous arrays", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "Did you know you can create an array of non-uniform data types, run it through a differential equation solver and still ensure there is no need for runtime type inference? We will dive into Julia's array interface and explore how it allows us to implement such unusual, yet useful, data structures in a type-stable manner.", "description": "This talk will discuss the concept of arrays in Julia and provide the motivation for using heterogenous arrays. The starting point is the base Julia `Array` type which will serve as the description of the \"ideal\" array. However, for many applications it is desirable to have types that are different from the \"ideal\" `Array`, but share enough properties that it can still be thought of as an array. For this Julia provides an `AbstractArray` supertype allowing custom types to use interfaces primarily written for `Array`. We will briefly discuss the properties of the \"ideal\" `Array`s with counterexamples from commonly used custom array types. Here, we arrive at the main focus of the talk; heterogenous arrays that are arrays where the elements are not of the same type.\r\n\r\nOne can take two different approaches to heterogenous arrays. The first and boldest one is to have an array of genuinely different data types such as integers and floating point numbers. The second approach is less radical and is concerned with arrays that store the same underlying data type, but where the data type serves as an abstraction affecting its interpretation and its operations.\r\n\r\nWe choose to focus on the second approach by presenting an example where the state of a system is represented with `Unitful` quantities, thus keeping track of the units of measurements. This system is represented by ordinary differential equations and we want to run it through an ODE solver while keeping track of the unit of measurements throughout the process. We will review existing solutions for this kind of problem and discuss their advantages and disadvantages.\r\n\r\nOne of the major challenges of heterogenous arrays is type stability. In order to run efficiently, Julia must know the types to operate on when compiling a function. When the array elements are not of the same type, this poses additional challenges related to type inference. The talk will take the listeners through the process of implementing a heterogenous array type which supports type-stable broadcasting. We will also look into how the package `DifferentialEquations` is written in order to deal with heterogenous arrays. Finally, remaining challenges on the subject are addressed, and suggestions for further directions and work are provided.", "recording_license": "", "do_not_record": false, "persons": [{"code": "BF8VNY", "name": "Jakob Peder Pettersen", "avatar": "https://pretalx.com/media/avatars/BF8VNY_V9yf5w4.webp", "biography": "I work as a postdoctorial reseach fellow at UiT the Arctic University of Norway where I work on the Centre for New Antibacterial Strategies (CANS). In 2023, I graduated with a PhD in Biotechnology from the Norwegian University of Science and Technology.\r\n\r\nMy academic experience is centered around systems biology topics such as microbiome analysis, microbial metabolic modeling and gene expression analyses. In addition, I am passionate about scientific software engineering practices and high performance computing.", "public_name": "Jakob Peder Pettersen", "guid": "e682a6b0-da90-5058-888b-91a0f6b034b4", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/BF8VNY/"}], "links": [{"title": "GitHub repo of example implementation", "url": "https://github.com/yaccos/COMBAT.jl/tree/JuliaCon", "type": "related"}], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/DJCV9D/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/DJCV9D/", "attachments": [{"title": "Poster file", "url": "/media/juliacon-local-paris-2025/submissions/DJCV9D/resources/_mi6AwbU.pdf", "type": "related"}]}, {"guid": "f945416a-f367-5861-80f8-764b2e4443a3", "code": "JJCCXU", "id": 71932, "logo": null, "date": "2025-10-03T10:54:00+02:00", "start": "10:54", "duration": "00:03", "room": "Coffee room", "slug": "juliacon-local-paris-2025-71932-duckdb-as-backend-to-build-optimization-models-in-jump-jl", "url": "https://pretalx.com/juliacon-local-paris-2025/talk/JJCCXU/", "title": "DuckDB as backend to build optimization models in JuMP.jl", "subtitle": "", "track": "Posters", "type": "Poster", "language": "en", "abstract": "We use DuckDB as a backend to generate JuMP variables and constraints using SQL. This enables memory-efficient slicing, reusable indices, and traceable constraints. In TulipaEnergyModel.jl, this reduced preprocessing time and memory usage by up to 50%, improving scalability and clarity.", "description": "Data-centric modelling in energy, logistics, and finance often starts with a familiar dilemma: the dataset lives happily in a column store, while the optimisation problems are represented as arrays and sparse matrices. In this talk, we share a pragmatic solution to bridge that gap, using DuckDB as a lightweight SQL backend to generate JuMP variables and constraints directly from database queries. Our workflow keeps the data where it is efficient\u2014inside DuckDB\u2014while producing **linearised indices** that JuMP processes quickly and efficiently. We decouple **index construction** from **model construction**.  This separation brings three concrete benefits:\r\n* **Memory-aware slicing.**  DuckDB streams the relevant rows, so the Julia heap never needs to see the full fact table.\r\n* **Reusable indices.**  Once built, the same index table feeds multiple model constraints.\r\n* **Transparent provenance.**  Each constraint row carries its original SQL, providing an inherent audit trail.\r\n\r\nWhen implemented in the package Tulipaenergymodel.jl, the DuckDB approach helped reduce preprocessing time and memory footprint by 40 - 50%, making the code cleaner and more scalable for large optimization problems.", "recording_license": "", "do_not_record": false, "persons": [{"code": "3RVYDP", "name": "Abel Soares Siqueira", "avatar": "https://pretalx.com/media/avatars/3RVYDP_RluDDhB.webp", "biography": "Research Software Engineer at the Netherlands eScience Center.", "public_name": "Abel Soares Siqueira", "guid": "d0ad7f04-4dab-598b-b3bd-7596d8e9c618", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/3RVYDP/"}, {"code": "MQNYYE", "name": "Diego Tejada", "avatar": null, "biography": null, "public_name": "Diego Tejada", "guid": "c6d00cf5-9c18-566f-9eb6-019b3d55c170", "url": "https://pretalx.com/juliacon-local-paris-2025/speaker/MQNYYE/"}], "links": [], "feedback_url": "https://pretalx.com/juliacon-local-paris-2025/talk/JJCCXU/feedback/", "origin_url": "https://pretalx.com/juliacon-local-paris-2025/talk/JJCCXU/", "attachments": []}]}}]}}}