Counting Reward Automata: Sample Efficient Reinforcement Learning Through the Exploitation of Reward Function Structure
This work addresses the challenge of sample-efficient reinforcement learning for a broad range of tasks, including those beyond regular languages, by exploiting reward function structure, representing a novel advancement rather than an incremental improvement.
The paper tackles the problem of limited expressiveness in reinforcement learning reward functions by introducing counting reward automata, which can model any reward function expressible as a formal language, including unrestricted grammars, and shows that this approach outperforms competing methods in sample efficiency, automaton complexity, and task completion.
We present counting reward automata-a finite state machine variant capable of modelling any reward function expressible as a formal language. Unlike previous approaches, which are limited to the expression of tasks as regular languages, our framework allows for tasks described by unrestricted grammars. We prove that an agent equipped with such an abstract machine is able to solve a larger set of tasks than those utilising current approaches. We show that this increase in expressive power does not come at the cost of increased automaton complexity. A selection of learning algorithms are presented which exploit automaton structure to improve sample efficiency. We show that the state machines required in our formulation can be specified from natural language task descriptions using large language models. Empirical results demonstrate that our method outperforms competing approaches in terms of sample efficiency, automaton complexity, and task completion.