KAGE-Bench: Fast Known-Axis Visual Generalization Evaluation for Reinforcement Learning

TL;DR

KAGE-Bench introduces a fast, systematic benchmark for evaluating visual generalization in reinforcement learning by isolating individual visual shifts using a novel environment and axis-factorized observations.

Contribution

The paper presents KAGE-Env, a JAX-native platform that factorizes visual axes, and KAGE-Bench, a benchmark suite for analyzing visual generalization in RL, enabling controlled and rapid evaluation.

Findings

Background and photometric shifts often cause agent failure.

Agent-appearance shifts are less impactful on performance.

Visual shifts can break task completion without affecting reward.

Abstract

Pixel-based reinforcement learning agents often fail under purely visual distribution shift even when latent dynamics and rewards are unchanged, but existing benchmarks entangle multiple sources of shift and hinder systematic analysis. We introduce KAGE-Env, a JAX-native 2D platformer that factorizes the observation process into independently controllable visual axes while keeping the underlying control problem fixed. By construction, varying a visual axis affects performance only through the induced state-conditional action distribution of a pixel policy, providing a clean abstraction for visual generalization. Building on this environment, we define KAGE-Bench, a benchmark of six known-axis suites comprising 34 train-evaluation configuration pairs that isolate individual visual shifts. Using a standard PPO-CNN baseline, we observe strong axis-dependent failures, with background and photometric shifts often collapsing success, while agent-appearance shifts are comparatively benign. Several shifts preserve forward motion while breaking task completion, showing that return alone can obscure generalization failures. Finally, the fully vectorized JAX implementation enables up to 33M environment steps per second on a single GPU, enabling fast and reproducible sweeps over visual factors. Code: https://avanturist322.github.io/KAGEBench/.