Reproduce “A Practical Method for Constructing Equivariant Multilayer Perceptrons for Arbitrary Matrix Groups” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Marc Finzi, Max Welling, Andrew Gordon Wilson
The paper presents EMLP, which constructs equivariant layers for finite-dimensional representations of real Lie groups with finitely many components. It turns discrete and infinitesimal generators into constraints and solves for linear maps satisfying them. On a three-dimensional double-spring-pendulum task, an O(2)-equivariant Hamiltonian neural network reports rollout relative error 0.012(2), versus 0.028 for an ordinary MLP; this comparison is limited to that setup.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.