AbideGym: Turning Static RL Worlds into Adaptive Challenges
This addresses the issue of static benchmarks in RL for researchers, though it is incremental as it builds on existing MiniGrid environments.
The paper tackles the problem of brittle reinforcement learning policies by introducing AbideGym, a dynamic MiniGrid wrapper that enforces intra-episode adaptation through agent-aware perturbations, resulting in a modular and reproducible evaluation framework for improving resilience.
Agents trained with reinforcement learning often develop brittle policies that fail when dynamics shift, a problem amplified by static benchmarks. AbideGym, a dynamic MiniGrid wrapper, introduces agent-aware perturbations and scalable complexity to enforce intra-episode adaptation. By exposing weaknesses in static policies and promoting resilience, AbideGym provides a modular, reproducible evaluation framework for advancing research in curriculum learning, continual learning, and robust generalization.