Institution profile

LAAS-CNRS

Academic institutioneurope · fr
Official website
Research library95linked papers
Opportunities0open roles
Selected work

Representative Papers

Online Stochastic Matching: A Polytope Perspective

Dec 29, 2021

This paper studies online stochastic matching under graph-theoretic compatibility constraints, where items of distinct classes arrive according to independent Poisson processes, compatibility is encoded by an undirected graph, and unmatched items are queued. Targeting the joint optimization of stability, matching delay, and long-run matching rate, we establish—for the first time—the fundamental connection between the existence of stable policies, the dimension of the convex polyhedron formed by nonnegative solutions to conservation equations, and the structural properties of the compatibility graph. We propose a novel policy design paradigm wherein performance bounds are characterized by the vertices of this polyhedron. Leveraging stochastic process modeling, graph theory, and convex analysis, we construct stable policies that either achieve or approximate these vertex bounds. These policies maximize the long-run matching rate while ensuring system stability and yield tight theoretical bounds on matching delay in terms of graph structure.

2 citationsRead paper

Reliability of Single-Level Equality-Constrained Inverse Optimal Control

Nov 22, 2024IEEE-RAS International Conference on Humanoid Robots

This study addresses the core problem in inverse optimal control of recovering cost function weights from human motion data. To overcome the low computational efficiency and poor noise robustness of existing bilevel optimization approaches, we propose a single-level reconstruction method based on optimality condition minimization. Our method equivalently reformulates the original bilevel optimization into a single-level problem, achieving significant speedup while strictly preserving modeling fidelity. Theoretical analysis and empirical evaluation demonstrate strong robustness to measurement noise. In numerical simulations of planar reaching tasks, the proposed method achieves a 15× speedup over classical bilevel algorithms and maintains stable convergence and accurate parameter estimation even under high noise levels. This work establishes a new paradigm for real-time, reliable human motion modeling and behavioral inference.

1 citationsRead paper
Recent publications

Latest Papers

An Explainable GNN Framework for Component-Level Anomaly Diagnosis

Aug 10, 2026

This work addresses a critical limitation in anomaly diagnosis for industrial systems, where existing approaches often misattribute anomalies to individual sensors while overlooking disruptions in the dynamic interactions among system components. To overcome this sensor-level attribution paradigm, the authors propose an interpretable graph neural network framework that treats anomalies as symptoms of altered influence relationships between components, thereby enabling root-cause localization at the component level. By explicitly modeling sensor dependencies and integrating explainability techniques, the method identifies key influence pathways and pinpoints the true faulty components. Experimental results demonstrate that the proposed approach significantly outperforms state-of-the-art methods in both diagnostic accuracy and interpretability, achieving precise identification and prioritization of faulty components within complex industrial systems.

0 citationsRead paper

Early Failure Prediction from Near-Anomaly Detection: A Proactive Approach

Jul 29, 2026

This study addresses the challenge that traditional anomaly detection methods struggle to identify samples near the boundary between normal and anomalous states, thereby failing to enable early fault warnings. To overcome this limitation, the paper introduces the novel concept of “near-anomalies” and proposes CANARI, an unsupervised method grounded in Christoffel function theory to model such borderline cases. By moving beyond conventional dual-threshold mechanisms, CANARI proactively identifies unlabeled samples likely to evolve into failures. Experimental results on both synthetic and real-world industrial printed circuit board in-circuit test data demonstrate that CANARI significantly outperforms existing baselines, offering a robust foundation for predictive maintenance and quality control.

0 citationsRead paper