Post #294

SE(3)-Transformers A blog post about a recent paper (NeurIPS 2020) that introduces group theory to set functions. It seems like it performs on par with state-of-the-art methods for classification and regression, but at least is provably equivariant.