Spatial Attention Kinetic Networks with E(n)-Equivariance

Yuanqing Wang, John Chodera

Abstract:

Neural networks that are equivariant to rotations, translations, reflections, and permutations on $n$-dimensional geometric space have shown promise in physical modeling for tasks such as accurately but inexpensively modeling complex potential energy surfaces to guiding the sampling of complex dynamical systems or forecasting their time evolution.
Current state-of-the-art methods employ spherical harmonics to encode higher-order interactions among particles, which are computationally expensive.
In this paper, we propose a simple alternative functional form that uses neurally parametrized linear combinations of edge vectors to achieve equivariance while still universally approximating node environments.
Incorporating this insight, we design \emph{spatial attention kinetic networks} with E(n)-equivariance, or SAKE, which are competitive in many-body system modeling tasks while being significantly faster.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning

TL;DR: Equivariant functional form termed spatial attention uses neurally parametrized linear combinations of edge vectors to equivariantly yet describe node environments

Community Implementations: [ 2 code implementations](https://www.catalyzex.com/paper/arxiv:2301.08893/code)