Maximum Causal Entropy Inverse Constrained Reinforcement Learning

TL;DR

This paper introduces a maximum causal entropy-based method for inverse constrained reinforcement learning that learns constraints from demonstrations, ensuring agents adhere to implicit environment norms and outperform existing approaches.

Contribution

The paper proposes a novel maximum causal entropy framework for inverse constrained reinforcement learning, with proven convergence and scalable approximation for complex environments.

Findings

Outperforms state-of-the-art methods across various tasks

Handles stochastic dynamics and continuous spaces

Effective transferability of learned cost functions

Abstract

When deploying artificial agents in real-world environments where they interact with humans, it is crucial that their behavior is aligned with the values, social norms or other requirements of that environment. However, many environments have implicit constraints that are difficult to specify and transfer to a learning agent. To address this challenge, we propose a novel method that utilizes the principle of maximum causal entropy to learn constraints and an optimal policy that adheres to these constraints, using demonstrations of agents that abide by the constraints. We prove convergence in a tabular setting and provide an approximation which scales to complex environments. We evaluate the effectiveness of the learned policy by assessing the reward received and the number of constraint violations, and we evaluate the learned cost function based on its transferability to other agents. Our method has been shown to outperform state-of-the-art approaches across a variety of tasks and environments, and it is able to handle problems with stochastic dynamics and a continuous state-action space.