HackAtari: Atari Learning Environments for Robust and Continual Reinforcement Learning

TL;DR

HackAtari introduces a framework for creating novel Atari game scenarios with controlled modifications to test and improve reinforcement learning agents' robustness and adaptability in the face of novelty and changing environments.

Contribution

The paper presents HackAtari, a novel framework for generating diverse, controlled Atari environments to evaluate and enhance RL agents' robustness and generalization capabilities.

Findings

Current RL agents show robustness failures in original environments.

HackAtari effectively creates varied scenarios for testing RL agents.

Using HackAtari improves RL agents' robustness and alignment.

Abstract

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.