07-31, 13:20–13:30 (UTC), Red Track
This talk will describe work done using the NeuralNetDiffEq & Flux packages, as part of the Google Code In 2019 program, including writing a second order ODE solver. I will also talk about pursuing Julia beyond GCI, including how I used Julia to create Corona-Net, a binary and multi-class symptoms localisation system for COVID-19. Through this talk, I hope share on how GCI with Julia has accelerated my progress in ML, and how Julia can be used to help contribute to the fight against COVID-19.
Some highlights for Google Code In were working with the NeuralNetDiffEq and Flux packages. In this talk, I will describe the process of writing a 2nd order ordinary differential equations solver using Neural Networks as part of GCI. Subsequently, I will cover how I used Flux to implement UNet's Fully-Convolutional Network architecture, for binary and multi-class (3 symptoms: ground glass, consolidation, pleural effusion) segmentation on the COVID-19 CT Segmentation Datasets, and discuss the Julia versus Python workflow in creating Corona-Net. Lastly, I will reflect on the impact of Google Code In Julia track on teenage ML & Open Source enthusiasts like myself, and how it ties in to my subsequent opportunities at NVIDIA.
(A big thank you to Chris Rackauckas, Kirill Zubov, Avik Sengupta, Dhairya Gandhi, Avik Pal and Logan Kilpatrick for all their guidance in GCI, JuliaCon and beyond!)
16 year old ML enthusiast, aspiring CV researcher, intern at NVIDIA's AI Tech Center