HackAtari: Atari Learning Environments for Robust and Continual Reinforcement Learning
This work addresses robustness and generalization issues in reinforcement learning agents, which is crucial for their effective deployment, though it appears incremental as it builds on existing benchmarks and methods.
The authors tackled the problem of reinforcement learning agents struggling with novel situations by proposing HackAtari, a framework that introduces controlled novelty to the Atari Learning Environment, and demonstrated that current agents have robustness failures while showing HackAtari can improve robustness and alignment in experiments with C51 and PPO.
Artificial agents' adaptability to novelty and alignment with intended behavior is crucial for their effective deployment. Reinforcement learning (RL) leverages novelty as a means of exploration, yet agents often struggle to handle novel situations, hindering generalization. To address these issues, we propose HackAtari, a framework introducing controlled novelty to the most common RL benchmark, the Atari Learning Environment. HackAtari allows us to create novel game scenarios (including simplification for curriculum learning), to swap the game elements' colors, as well as to introduce different reward signals for the agent. We demonstrate that current agents trained on the original environments include robustness failures, and evaluate HackAtari's efficacy in enhancing RL agents' robustness and aligning behavior through experiments using C51 and PPO. Overall, HackAtari can be used to improve the robustness of current and future RL algorithms, allowing Neuro-Symbolic RL, curriculum RL, causal RL, as well as LLM-driven RL. Our work underscores the significance of developing interpretable in RL agents.