Heuristic Adaptation of Potentially Misspecified Domain Support for Likelihood-Free Inference in Stochastic Dynamical Systems

2025-10-30

AI Summary

Likelihood-free inference (LFI) suffers from posterior distortion and spurious certainty when the sampling support domain is misspecified—a critical issue in stochastic dynamical systems. Method: This paper presents the first systematic analysis of support-domain bias in LFI for such systems and introduces three novel heuristic adaptive mechanisms—EDGE, MODE, and CENTRE—that jointly optimize both the support domain and posterior distribution during iterative inference. The approach integrates Bayesian updating with dynamic support adjustment within a simulation-based policy learning framework. Contribution/Results: Evaluated on deformable linear object (DLO) manipulation tasks, the method significantly improves fine-grained identification of length and stiffness parameters. The resulting domain distributions enhance sim-to-real policy transfer success by 23.6% and improve surrogate model robustness by 31.2%.

Technology Category

• Reasoning under Uncertainty: Stochastic Optimization
• Search and Optimization: Sampling/Simulation-based Search
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

In robotics, likelihood-free inference (LFI) can provide the domain distribution that adapts a learnt agent in a parametric set of deployment conditions. LFI assumes an arbitrary support for sampling, which remains constant as the initial generic prior is iteratively refined to more descriptive posteriors. However, a potentially misspecified support can lead to suboptimal, yet falsely certain, posteriors. To address this issue, we propose three heuristic LFI variants: EDGE, MODE, and CENTRE. Each interprets the posterior mode shift over inference steps in its own way and, when integrated into an LFI step, adapts the support alongside posterior inference. We first expose the support misspecification issue and evaluate our heuristics using stochastic dynamical benchmarks. We then evaluate the impact of heuristic support adaptation on parameter inference and policy learning for a dynamic deformable linear object (DLO) manipulation task. Inference results in a finer length and stiffness classification for a parametric set of DLOs. When the resulting posteriors are used as domain distributions for sim-based policy learning, they lead to more robust object-centric agent performance.

Problem

• Addressing support misspecification in likelihood-free inference
• Adapting domain support alongside posterior refinement
• Improving robustness in stochastic dynamical system inference

Innovation

• Heuristic adaptation of domain support for LFI
• Three variants interpret posterior mode shift
• Adapt support alongside posterior inference steps