Reward learning from human preferences and demonstrations in Atari
This addresses the challenge of specifying rewards for complex tasks in reinforcement learning, though it is incremental as it builds on existing methods for human feedback.
The paper tackled the problem of training reinforcement learning agents without manually specified rewards by combining human demonstrations and preferences to learn a reward function, achieving superhuman performance on 2 out of 9 Atari games and outperforming imitation learning in 7 games.
To solve complex real-world problems with reinforcement learning, we cannot rely on manually specified reward functions. Instead, we can have humans communicate an objective to the agent directly. In this work, we combine two approaches to learning from human feedback: expert demonstrations and trajectory preferences. We train a deep neural network to model the reward function and use its predicted reward to train an DQN-based deep reinforcement learning agent on 9 Atari games. Our approach beats the imitation learning baseline in 7 games and achieves strictly superhuman performance on 2 games without using game rewards. Additionally, we investigate the goodness of fit of the reward model, present some reward hacking problems, and study the effects of noise in the human labels.