Generalized Physics-Informed Learning through Language-Wide Differentiable Programming by Rackauckas
Chris Rackauckas (MIT), "Generalized Physics-Informed Learning through Language-Wide Differentiable Programming"
Scientific computing is increasingly incorporating the advancements in machine learning to allow for data-driven physics-informed modeling approaches. However, re-targeting existing scientific computing workloads to machine learning frameworks is both costly and limiting, as scientific simulations tend to use the full feature set of a general purpose programming language. In this manuscript we develop an infrastructure for incorporating deep learning into existing scientific computing code through Differentiable Programming (∂P). We describe a ∂P system that is able to take gradients of full Julia programs, making Automatic Differentiation a first class language feature and compatibility with deep learning pervasive. Our system utilizes the one-language nature of Julia package development to augment the existing package ecosystem with deep learning, supporting almost all language constructs (control flow, recursion, mutation, etc.) while generating high-performance code without requiring any user intervention or refactoring to stage computations. We showcase several examples of physics-informed learning which directly utilizes this extension to existing simulation code: neural surrogate models, machine learning on simulated quantum hardware, and data-driven stochastic dynamical model discovery with neural stochastic differential equations.
Code is available at https://github.com/MikeInnes/zygote-paper
AAAI 2020 Spring Symposium on Combining Artificial Intelligence and Machine Learning with Physics Sciences, March 23-25, 2020 (https://sites.google.com/view/aaai-mlps)
Papers: https://sites.google.com/view/aaai-mlps/proceedings
Slides: https://sites.google.com/view/aaai-mlps/program
Видео Generalized Physics-Informed Learning through Language-Wide Differentiable Programming by Rackauckas канала MLPS - Combining AI and ML with Physics Sciences
Scientific computing is increasingly incorporating the advancements in machine learning to allow for data-driven physics-informed modeling approaches. However, re-targeting existing scientific computing workloads to machine learning frameworks is both costly and limiting, as scientific simulations tend to use the full feature set of a general purpose programming language. In this manuscript we develop an infrastructure for incorporating deep learning into existing scientific computing code through Differentiable Programming (∂P). We describe a ∂P system that is able to take gradients of full Julia programs, making Automatic Differentiation a first class language feature and compatibility with deep learning pervasive. Our system utilizes the one-language nature of Julia package development to augment the existing package ecosystem with deep learning, supporting almost all language constructs (control flow, recursion, mutation, etc.) while generating high-performance code without requiring any user intervention or refactoring to stage computations. We showcase several examples of physics-informed learning which directly utilizes this extension to existing simulation code: neural surrogate models, machine learning on simulated quantum hardware, and data-driven stochastic dynamical model discovery with neural stochastic differential equations.
Code is available at https://github.com/MikeInnes/zygote-paper
AAAI 2020 Spring Symposium on Combining Artificial Intelligence and Machine Learning with Physics Sciences, March 23-25, 2020 (https://sites.google.com/view/aaai-mlps)
Papers: https://sites.google.com/view/aaai-mlps/proceedings
Slides: https://sites.google.com/view/aaai-mlps/program
Видео Generalized Physics-Informed Learning through Language-Wide Differentiable Programming by Rackauckas канала MLPS - Combining AI and ML with Physics Sciences
Показать
Комментарии отсутствуют
Информация о видео
30 марта 2020 г. 20:43:01
00:54:15
Другие видео канала
![Accelerating Simulation of Stiff Nonlinear Systems using Continuous-Time Echo State Nets, Rackauckas](https://i.ytimg.com/vi/3lM-Stc7z28/default.jpg)
![Differentiable Programming - An exciting generalisation of deep neural networks by Dr Viral B. Shah](https://i.ytimg.com/vi/6V75lblJ2ns/default.jpg)
![Top 5 Programming Languages to Learn in 2020 to Get a Job Without a College Degree](https://i.ytimg.com/vi/tCAt8eEKPDc/default.jpg)
![How to Learn Faster with the Feynman Technique (Example Included)](https://i.ytimg.com/vi/_f-qkGJBPts/default.jpg)
![Physics-Informed Neural Networks (PINNs) - Chris Rackauckas | Podcast #42](https://i.ytimg.com/vi/OmySUTFwh2g/default.jpg)
![Google's Deep Mind Explained! - Self Learning A.I.](https://i.ytimg.com/vi/TnUYcTuZJpM/default.jpg)
![Machine Learning for Fluid Dynamics: Models and Control](https://i.ytimg.com/vi/gb_C9LcjDSI/default.jpg)
![Map of Computer Science](https://i.ytimg.com/vi/SzJ46YA_RaA/default.jpg)
![How language shapes the way we think | Lera Boroditsky](https://i.ytimg.com/vi/RKK7wGAYP6k/default.jpg)
![Machine Learning Zero to Hero (Google I/O'19)](https://i.ytimg.com/vi/VwVg9jCtqaU/default.jpg)
![Surrogate Model Based Optimization - Yong Hoon Lee](https://i.ytimg.com/vi/SdRf7kRhip0/default.jpg)
![Differentiable Programming with Julia by Mike Innes](https://i.ytimg.com/vi/LjWzgTPFu14/default.jpg)
![Zico Kolter - Incorporating physics and decision making into deep learning via implicit layers](https://i.ytimg.com/vi/3K9Et7t1XT8/default.jpg)
![Why We Struggle Learning Languages | Gabriel Wyner | TEDxNewBedford](https://i.ytimg.com/vi/iBMfg4WkKL8/default.jpg)
![The secrets of learning a new language | Lýdia Machová](https://i.ytimg.com/vi/o_XVt5rdpFY/default.jpg)
![Differentiable Physics (for Deep Learning), Overview Talk by Nils Thuerey](https://i.ytimg.com/vi/BwuRTpTR2Rg/default.jpg)
![Stephan Hoyer: "Improving PDE solvers and PDE-constrained optimization with deep learning and di..."](https://i.ytimg.com/vi/ysIr_csR6gA/default.jpg)
![Neural Network Architectures & Deep Learning](https://i.ytimg.com/vi/oJNHXPs0XDk/default.jpg)
![Targeted use of deep learning for physics-informed model discovery by Nathan Kutz](https://i.ytimg.com/vi/oS7NWbSe000/default.jpg)
![What's Hidden in a Randomly Weighted Neural Network?](https://i.ytimg.com/vi/C6Tj8anJO-Q/default.jpg)