Benchmarking Batch Deep Reinforcement Learning Algorithms
This work addresses the problem of evaluating batch RL algorithms for researchers, providing a strong baseline but is incremental in nature.
The paper benchmarks recent batch deep reinforcement learning algorithms on Atari using data from a partially-trained policy, finding that many underperform compared to online DQN and the behavioral policy, and introduces a modified Batch-Constrained Q-learning algorithm that outperforms existing methods.
Widely-used deep reinforcement learning algorithms have been shown to fail in the batch setting--learning from a fixed data set without interaction with the environment. Following this result, there have been several papers showing reasonable performances under a variety of environments and batch settings. In this paper, we benchmark the performance of recent off-policy and batch reinforcement learning algorithms under unified settings on the Atari domain, with data generated by a single partially-trained behavioral policy. We find that under these conditions, many of these algorithms underperform DQN trained online with the same amount of data, as well as the partially-trained behavioral policy. To introduce a strong baseline, we adapt the Batch-Constrained Q-learning algorithm to a discrete-action setting, and show it outperforms all existing algorithms at this task.