Reproduce “Covariance in Physics and Convolutional Neural Networks” 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
Miranda C. N. Cheng, Vassilis Anagiannis, Maurice Weiler, Pim de Haan, Taco S. Cohen, Max Welling
The authors derive the form of a convolution for feature fields on a manifold from linearity, locality, covariance, and weight sharing. The construction transports nearby features along geodesics into a common local frame and accounts for the metric’s volume factor. The result is conditional on the stated geometric and locality assumptions; the paper reports no experiments or performance gains.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.