More Efficient Exploration with Symbolic Priors on Action Sequence Equivalences
arXiv:2110.10632
Abstract
Incorporating prior knowledge in reinforcement learning algorithms is mainly an open question. Even when insights about the environment dynamics are available, reinforcement learning is traditionally used in a tabula rasa setting and must explore and learn everything from scratch. In this paper, we consider the problem of exploiting priors about action sequence equivalence: that is, when different sequences of actions produce the same effect. We propose a new local exploration strategy calibrated to minimize collisions and maximize new state visitations. We show that this strategy can be computed at little cost, by solving a convex optimization problem. By replacing the usual epsilon-greedy strategy in a DQN, we demonstrate its potential in several environments with various dynamic structures.
References in corpus (8)
- Provably Efficient Maximum Entropy Exploration
- Efficient Exploration via State Marginal Matching
- Never Give Up: Learning Directed Exploration Strategies
- Is Deep Reinforcement Learning Really Superhuman on Atari? Leveling the playing field
- Symmetry Learning for Function Approximation in Reinforcement Learning
- Geometric Entropic Exploration
- Monte-Carlo Graph Search for AlphaZero
- On the Sensory Commutativity of Action Sequences for Embodied Agents