Model Predictive Control

Equivariant Action Sampling for Reinforcement Learning and Planning

We study the problem of action sampling and propose a method to incorporate equivariance properties to the action sampling procedure.