Time-Myopic Go-Explore: Learning A State Representation for the Go-Explore Paradigm
This work addresses the detachment problem in Go-Explore for reinforcement learning in sparse-reward environments, offering a learned representation that improves exploration, though it is incremental as it builds on an existing paradigm.
The paper tackles the problem of exploring large state spaces with sparse rewards in reinforcement learning by introducing a time-myopic state representation that clusters temporally close states and predicts time between them, adapting it to the Go-Explore paradigm; it shows improved performance on hard exploration Atari environments like MontezumaRevenge, Gravitar, and Frostbite, providing a more general alternative to hand-crafted heuristics.
Very large state spaces with a sparse reward signal are difficult to explore. The lack of a sophisticated guidance results in a poor performance for numerous reinforcement learning algorithms. In these cases, the commonly used random exploration is often not helpful. The literature shows that this kind of environments require enormous efforts to systematically explore large chunks of the state space. Learned state representations can help here to improve the search by providing semantic context and build a structure on top of the raw observations. In this work we introduce a novel time-myopic state representation that clusters temporal close states together while providing a time prediction capability between them. By adapting this model to the Go-Explore paradigm (Ecoffet et al., 2021b), we demonstrate the first learned state representation that reliably estimates novelty instead of using the hand-crafted representation heuristic. Our method shows an improved solution for the detachment problem which still remains an issue at the Go-Explore Exploration Phase. We provide evidence that our proposed method covers the entire state space with respect to all possible time trajectories without causing disadvantageous conflict-overlaps in the cell archive. Analogous to native Go-Explore, our approach is evaluated on the hard exploration environments MontezumaRevenge, Gravitar and Frostbite (Atari) in order to validate its capabilities on difficult tasks. Our experiments show that time-myopic Go-Explore is an effective alternative for the domain-engineered heuristic while also being more general. The source code of the method is available on GitHub.