Parseval Regularization for Continual Reinforcement Learning

Plasticity loss, trainability loss, and primacy bias have been identified as issues arising when training deep neural networks on sequences of tasks---referring to the increased difficulty in training on new tasks. We propose to use Parseval regularization, which maintains orthogonality of weight matrices, to preserve useful optimization properties and improve training in a continual reinforcement learning setting. We show that it provides significant benefits to RL agents on a suite of gridworld, CARL and MetaWorld tasks. We conduct comprehensive ablations to identify the source of its benefits and investigate the effect of certain metrics associated to network trainability including weight matrix rank, weight norms and policy entropy.

Authors

• Wesley Chung
• Lynn Cherif
• David Meger McGill University Samsung AI Center Montreal Mila - Quebec AI Institute
• Doina Precup McGill University Mila

Keywords

No keywords are indexed for this paper.