Go Beyond Imagination: Maximizing Episodic Reachability with World Models
This work addresses the problem of reward sparsity in reinforcement learning for researchers and practitioners, offering an incremental improvement over existing intrinsic reward methods.
The paper tackles efficient exploration in reinforcement learning for sparse reward tasks by introducing GoBI, a new intrinsic reward design that combines lifelong novelty with episodic reachability expansion using world models. The method significantly outperforms previous state-of-the-art methods on 12 challenging Minigrid navigation tasks and improves sample efficiency on DeepMind Control Suite locomotion tasks.
Efficient exploration is a challenging topic in reinforcement learning, especially for sparse reward tasks. To deal with the reward sparsity, people commonly apply intrinsic rewards to motivate agents to explore the state space efficiently. In this paper, we introduce a new intrinsic reward design called GoBI - Go Beyond Imagination, which combines the traditional lifelong novelty motivation with an episodic intrinsic reward that is designed to maximize the stepwise reachability expansion. More specifically, we apply learned world models to generate predicted future states with random actions. States with more unique predictions that are not in episodic memory are assigned high intrinsic rewards. Our method greatly outperforms previous state-of-the-art methods on 12 of the most challenging Minigrid navigation tasks and improves the sample efficiency on locomotion tasks from DeepMind Control Suite.