Trust Region Reward Optimization and Proximal Inverse Reward Optimization Algorithm

Inverse Reinforcement Learning (IRL) learns a reward function to explain expert demonstrations. Modern IRL methods often use the adversarial (minimax) formulation that alternates between reward and policy optimization, which often lead to unstable training. Recent non-adversarial IRL approaches improve stability by jointly learning reward and policy via energy-based formulations but lack formal guarantees. This work bridges this gap. We first present a unified view showing canonical non-adversarial methods explicitly or implicitly maximize the likelihood of expert behavior, which is equivalent to minimizing the expected return gap. This insight leads to our main contribution: Trust Region Reward Optimization (TRRO), a framework that guarantees monotonic improvement in this likelihood via a Minorization-Maximization process. We instantiate TRRO into Proximal Inverse Reward Optimization (PIRO), a practical and stable IRL algorithm. Theoretically, TRRO provides the IRL counterpart to the stability guarantees of Trust Region Policy Optimization (TRPO) in forward RL. Empirically, PIRO matches or surpasses state-of-the-art baselines in reward recovery, policy imitation with high sample efficiency on MuJoCo and Gym-Robotics benchmarks and a real-world animal behavior modeling task. 1 * Title used at submission and review: PIRO: Toward Stable Reward Learning for Inverse RL via Monotonic Policy Divergence Reduction.

† Main contributors. Yang Chen developed the theorems, completed the proofs, wrote the paper, and implemented the initial version of the algorithm. Menglin Zou led the experimental evaluation. Jiaqi Zhang and Junyi Yang validated the algorithm using toy models. Yitan Zhang conducted the experiments on robotics and animal behavior modeling tasks. The remaining authors contributed through critical discussions and feedback.