Incorporating Symmetry into Deep Dynamics Models for Improved Generalization
About
Recent work has shown deep learning can accelerate the prediction of physical dynamics relative to numerical solvers. However, limited physical accuracy and an inability to generalize under distributional shift limit its applicability to the real world. We propose to improve accuracy and generalization by incorporating symmetries into convolutional neural networks. Specifically, we employ a variety of methods each tailored to enforce a different symmetry. Our models are both theoretically and experimentally robust to distributional shift by symmetry group transformations and enjoy favorable sample complexity. We demonstrate the advantage of our approach on a variety of physical dynamics including Rayleigh B\'enard convection and real-world ocean currents and temperatures. Compared with image or text applications, our work is a significant step towards applying equivariant neural networks to high-dimensional systems with complex dynamics. We open-source our simulation, data, and code at \url{https://github.com/Rose-STL-Lab/Equivariant-Net}.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Operator learning | 1-D Burgers resolution 1024 (ID) | ID Error2.93 | 5 | |
| Solving 2-D Burgers equation | 2-D Burgers In-Distribution resolution 64^2 (test) | ID Error0.0261 | 5 | |
| Operator learning | 1-D Burgers resolution 1024 to 2048 (OOD) | OOD Error0.3222 | 5 | |
| Solving 2-D Burgers equation | 2-D Burgers Out-of-Distribution resolution 64^2 to 128^2 shift (test) | OOD Error0.2035 | 5 |