07-10, 17:40–18:40 (Europe/Amsterdam), REPL (2, main stage)
Current solutions to key 21st-century challenges are fundamentally limited by the functional properties of known materials. Designing new materials increasingly relies on computational modeling leading to research questions across multiple scales and scientific fields. While we believe Julia is well-poised to aid with tackling these challenges, Julia tools cannot simply duplicate the many well-established software solutions. We discuss challenges for Julia and outline opportunities going forward
Current solutions to key 21st-century challenges (such as climate change, food insecurity, healthcare, and communications) are fundamentally limited by the functional properties of known materials. Designing new materials increasingly relies on computational modeling with state-of-the-art workflows frequently interweaving first-principles with empirical modeling as well as data-driven approaches. To make advances, we often must establish novel connections across fields such as physics, chemistry, computer science, and applied math. The Julia language and user community are well-poised to address these challenges. However, with many well-established software solutions and their respective user communities already in existence, Julia tools cannot just duplicate existing functionality. Rather, it is crucial to integrate and add value to existing (often monolithic) codebases and software ecosystems. In this talk, we will discuss existing efforts in this area --- both by ourselves as well as the broader JuliaMolSim community --- and outline remaining challenges and opportunities going forward.
Michael is currently a tenure-track assistant professor for mathematics and materials science at EPFL, Lausanne. There he leads the MatMat research group, which aims to understand simulation error and improve robustness of materials modelling schemes. As part of this effort his group co-develops the density-functional toolkit.