ViSaRL: Visual Reinforcement Learning Guided by Human Saliency

arXiv:2403.10940v1 Announce Type: cross
Abstract: Training robots to perform complex control tasks from high-dimensional pixel input using reinforcement learning (RL) is sample-inefficient, because image observations are comprised primarily of task-irrelevant information. By contrast, humans are able to visually attend to task-relevant objects and areas. Based on this insight, we introduce Visual Saliency-Guided Reinforcement Learning (ViSaRL). Using ViSaRL to learn visual representations significantly improves the success rate, sample efficiency, and generalization of an RL agent on diverse tasks including DeepMind Control …

abstract arxiv contrast control cs.lg cs.ro human humans image information insight objects pixel reinforcement reinforcement learning robots sample tasks training type visual