Confidence Aware Inverse Constrained Reinforcement Learning

TL;DR

This paper introduces a confidence-aware inverse constrained reinforcement learning method that estimates constraints from expert demonstrations with a specified confidence level, enabling more reliable and informed decision-making in complex real-world problems.

Contribution

It proposes a novel ICRL approach that incorporates confidence levels, allowing users to specify desired confidence and determine if more data is needed.

Findings

Provides a method to estimate constraints with user-specified confidence.

Enables detection of insufficient expert data for reliable constraint learning.

Supports simultaneous learning of constraints and policies with confidence guarantees.

Abstract

In coming up with solutions to real-world problems, humans implicitly adhere to constraints that are too numerous and complex to be specified completely. However, reinforcement learning (RL) agents need these constraints to learn the correct optimal policy in these settings. The field of Inverse Constraint Reinforcement Learning (ICRL) deals with this problem and provides algorithms that aim to estimate the constraints from expert demonstrations collected offline. Practitioners prefer to know a measure of confidence in the estimated constraints, before deciding to use these constraints, which allows them to only use the constraints that satisfy a desired level of confidence. However, prior works do not allow users to provide the desired level of confidence for the inferred constraints. This work provides a principled ICRL method that can take a confidence level with a set of expert demonstrations and outputs a constraint that is at least as constraining as the true underlying constraint with the desired level of confidence. Further, unlike previous methods, this method allows a user to know if the number of expert trajectories is insufficient to learn a constraint with a desired level of confidence, and therefore collect more expert trajectories as required to simultaneously learn constraints with the desired level of confidence and a policy that achieves the desired level of performance.