Multi-Objective Constraint Inference using Inverse reinforcement learning

2026-05-07

AI Summary

Existing approaches struggle to effectively handle shared constraints and individual preferences inherent in multi-objective, heterogeneous expert behaviors. To address this challenge, this work proposes the MOCI framework, which, for the first time, leverages inverse reinforcement learning to simultaneously infer safety constraints and personalized objectives from heterogeneous expert trajectories. By integrating trajectory modeling with preference disentanglement techniques, MOCI achieves enhanced expressiveness and generalization capability in capturing conflicting behavioral patterns. Evaluated on standard grid-world benchmarks, the proposed method significantly outperforms existing approaches in both prediction accuracy and computational efficiency.

Abstract

Constraint inference is widely considered essential to align reinforcement learning agents with safety boundaries and operational guidelines by observing expert demonstrations. However, existing approaches typically assume homogeneous demonstrations (i.e., generated by a single expert or multiple experts with identical objectives). They also have limited ability to capture individual preferences and often suffer from computational inefficiencies. In this paper, we introduce Multi-Objective Constraint Inference (MOCI), a novel framework designed to jointly extract shared constraints and individual preferences from heterogeneous expert trajectories, where multiple experts pursue different objectives. MOCI effectively models and learns from diverse, and potentially conflicting, behaviors. Empirical evaluations demonstrate that MOCI significantly outperforms existing baselines, achieving improved predictive performance, and maintaining competitive computational efficiency on a standard grid-world benchmark. These results establish MOCI as an accurate, flexible, and computationally practical approach for real-world constraint inference and preference learning tasks.

Problem

• Constraint Inference
• Inverse Reinforcement Learning
• Multi-Objective
• Heterogeneous Demonstrations
• Preference Learning

Innovation

• Multi-Objective Constraint Inference
• Inverse Reinforcement Learning
• Heterogeneous Demonstrations
• Preference Learning
• Shared Constraints