Sequence Alignment

Align-RUDDER -- Learning From Few Demonstrations by Reward Redistribution

We generalise steerable E(3) equivariant graph neural networks such that node and edge updates are able to leverage covariant information. Published at ICLR 2022 (Oral).