Average Reward Reinforcement Learning for Omega-Regular and Mean-Payoff Objectives

1. 1Specify behavioral requirements in omega-regular languages
2. 2Compile these requirements into learning objectives
3. 3Use average reward reinforcement learning to optimize agent behavior
4. 4Evaluate the performance of the learned policy using mean-payoff objectives

Machine learning researchers and engineers working on reinforcement learning and formal verification can benefit from this research to improve agent behavior and automate reward function design

💡 Omega-regular languages can be used to specify behavioral requirements and automatically compile them into learning objectives for reinforcement learning

🤖 Average reward RL for omega-regular & mean-payoff objectives 📈