LOSS OF PLASTICITY IN CONTINUAL DEEP REINFORCEMENT LEARNING

In this paper, we characterize the behavior of canonical value-based deep reinforcement learning (RL) approaches under varying degrees of non-stationarity. In particular, we demonstrate that deep RL agents lose their ability to learn good policies when they cycle through a sequence of Atari 2600 games. This phenomenon is alluded to in prior work under various guises-e.g., loss of plasticity, implicit under-parameterization, primacy bias, and capacity loss. We investigate this phenomenon closely at scale and analyze how the weights, gradients, and activations change over time in several experiments with varying experimental conditions (e.g., similarity between games, number of games, number of frames per game). Our analysis shows that the activation footprint of the network becomes sparser, contributing to the diminishing gradients. We investigate a remarkably simple mitigation strategy-Concatenated ReLUs (CReLUs) activation function-and demonstrate its effectiveness in facilitating continual learning in a changing environment.