Active Constraint Learning in High Dimensions from Demonstrations

2025/12/28 by Zheng Qiu, Qiu, Zheng, Chih-Yuan Chiu +3
Computer Science · #Gaussian Processes and Bayesian Inference #Machine Learning and Algorithms #Bayesian Modeling and Causal Inference

paper · doi:10.48550/arxiv.2512.22757

Abstract

We present an iterative active constraint learning (ACL) algorithm, within the learning from demonstrations (LfD) paradigm, which intelligently solicits informative demonstration trajectories for inferring an unknown constraint in the demonstrator's environment. Our approach iteratively trains a Gaussian process (GP) on the available demonstration dataset to represent the unknown constraints, uses the resulting GP posterior to query start/goal states, and generates informative demonstrations which are added to the dataset. Across simulation and hardware experiments using high-dimensional nonlinear dynamics and unknown nonlinear constraints, our method outperforms a baseline, random-sampling based method at accurately performing constraint inference from an iteratively generated set of sparse but informative demonstrations.

Citations

• Bayesian Constraint Inference from User Demonstrations Based on Margin-Respecting Preference Models
• Intent Demonstration in General-Sum Dynamic Games via Iterative Linear-Quadratic Approximations
• Constraint Inference in Control Tasks from Expert Demonstrations via Inverse Optimization
• Gaussian Process Constraint Learning for Scalable Chance-Constrained Motion Planning from Demonstrations
• PEBBLE: Feedback-Efficient Interactive Reinforcement Learning via Relabeling Experience and Unsupervised Pre-training
• Maximum Likelihood Constraint Inference from Stochastic Demonstrations
• Uncertainty-Aware Constraint Learning for Adaptive Safe Motion Planning from Demonstrations
• Learning Constraints from Locally-Optimal Demonstrations under Cost Function Uncertainty
• Learning Parametric Constraints in High Dimensions from Demonstrations
• Risk-sensitive Inverse Reinforcement Learning via Semi- and\n Non-Parametric Methods
• Learning how to Active Learn: A Deep Reinforcement Learning Approach
• On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming

Related

• Learning Constraint Network from Demonstrations via Positive-Unlabeled Learning with Memory Replay
• Learning Constraints from Demonstrations
• Learning Constraints from Locally-Optimal Demonstrations under Cost Function Uncertainty
• Learning Parametric Constraints in High Dimensions from Demonstrations
• Learning Constraints from Stochastic Partially-Observed Closed-Loop Demonstrations