Abstract: This talk will give an overview of how deep learning can be combined with traditional numerical methods to create improved methods for scientific computing. I will highlight two recent examples from research: using deep learning to improve discretizations for solving partial differential equations , and using deep learning to reparameterize optimization landscapes for PDE constrained structural optimization . I will also briefly introduce JAX , an open source library from Google for composable transformations of Python/NumPy programs, including automatic differentiation, vectorization and JIT compilation for accelerators. JAX is particularly suitable for scientific applications, including hybrid machine learning / simulation codes. Full details about this meeting will be posted here: https://bids.github.io/MLStatsForum/.
- Bar-Sinai*, Y., Hoyer*, S., Hickey, J. & Brenner, M. P. Learning data-driven discretizations for partial differential equations. Proceedings of the National Academy of Sciences 201814058 (2019). doi:10.1073/pnas.1814058116
- Hoyer, S., Sohl-Dickstein, J. & Greydanus, S. Neural reparameterization improves structural optimization. arXiv [cs.LG] (2019). https://arxiv.org/abs/1909.04240
The Machine Learning and Science Forum (formerly the Berkeley Statistics and Machine Learning Forum) meets biweekly to discuss current applications across a wide variety of research domains in the physical sciences and beyond. Hosted by UC Berkeley Physics Professor and BIDS Senior Fellow Uros Seljak, these active sessions bring together domain scientists, statisticians, and computer scientists who are either developing state-of-the-art methods or are interested in applying these methods in their research. To receive email notifications about upcoming meetings, or to request more information, please contact firstname.lastname@example.org. All interested members of the UC Berkeley and Berkeley Lab communities are welcome and encouraged to attend.
Stephan Hoyer is a physicist, data scientist and software engineer, and currently, a software engineer and researcher at Google on the Accelerated Sciences team, which leverages Google’s unique expertise to solve scientific research problems, e.g. deep learning for drug discovery. Previously, he was a research scientist at The Climate Corporation, where I worked on models of weather and climate risk for agriculture. At Climate, he wrote xarray, a library for labeled arrays in Python. He is also a frequent contributor to the open source scientific Python stack. He has a PhD in Physics from UC Berkeley (2013), and a B.A. in Physics from Swarthmore College (2008). Read more at stephanhoyer.com.