Maximum Causal Entropy Inverse Constrained Reinforcement Learning
This addresses the challenge of specifying implicit constraints for safe and aligned AI deployment in real-world settings, representing a novel method for a known bottleneck.
The paper tackles the problem of aligning AI agents with implicit constraints in human-interactive environments by proposing a method that uses maximum causal entropy to learn constraints and an optimal policy from demonstrations, achieving superior performance in reward and constraint violations across various tasks.
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.