MaxInfoRL: Boosting exploration in reinforcement learning through information gain maximization
This work addresses the problem of effective exploration in RL for researchers and practitioners, offering a novel method that improves performance in complex scenarios, though it builds on existing intrinsic reward concepts.
The paper tackles the challenge of balancing exploration and exploitation in reinforcement learning by introducing MaxInfoRL, a framework that directs exploration towards informative transitions using intrinsic rewards like information gain, and demonstrates that it achieves sublinear regret in multi-armed bandits and superior performance in hard exploration and visual control tasks.
Reinforcement learning (RL) algorithms aim to balance exploiting the current best strategy with exploring new options that could lead to higher rewards. Most common RL algorithms use undirected exploration, i.e., select random sequences of actions. Exploration can also be directed using intrinsic rewards, such as curiosity or model epistemic uncertainty. However, effectively balancing task and intrinsic rewards is challenging and often task-dependent. In this work, we introduce a framework, MaxInfoRL, for balancing intrinsic and extrinsic exploration. MaxInfoRL steers exploration towards informative transitions, by maximizing intrinsic rewards such as the information gain about the underlying task. When combined with Boltzmann exploration, this approach naturally trades off maximization of the value function with that of the entropy over states, rewards, and actions. We show that our approach achieves sublinear regret in the simplified setting of multi-armed bandits. We then apply this general formulation to a variety of off-policy model-free RL methods for continuous state-action spaces, yielding novel algorithms that achieve superior performance across hard exploration problems and complex scenarios such as visual control tasks.