Institution profile

Institut de Recherche en Informatique de Toulouse

Academic institutioneurope · fr
Official website
Research library124linked papers
Opportunities0open roles
Selected work

Representative Papers

Implementación de una Base de Datos Relacional Difusa. Un Caso en la Industria del Cartón

Feb 13, 2025Rev. Colomb. de Computación

To address the uncertainty modeling challenge arising from coexisting perceptual fuzzy attributes (e.g., visually/tactually perceived defects, thickness uniformity) and precise measurement data in the coating process of corrugated board manufacturing in the Maule region of Chile, this paper proposes an extended relational database solution based on the GEFRED fuzzy data model. It represents the first industrial deployment of the GEFRED model in South American corrugated board production lines. Implemented as a customized PostgreSQL extension, the solution integrates fuzzy set theory, possibility distribution modeling, and unified management of mixed-precision data, while supporting fuzzy SQL queries and rule-driven real-time quality alerts. A prototype system deployed in the conversion department achieves 92.7% accuracy in fuzzy query execution and improves quality decision consistency by 38%, thereby advancing the engineering practice of uncertain data governance in manufacturing.

6 citationsRead paper

Efficient Robust Conformal Prediction via Lipschitz-Bounded Networks

Jun 05, 2025

Traditional conformal prediction (CP) fails under adversarial attacks, while existing robust CP methods suffer from excessively large prediction sets or high computational overhead on large-scale tasks. To address this, we propose Lip-RCP—the first efficient robust prediction framework that deeply integrates 1-Lipschitz robust neural networks with CP. Methodologically, we impose Lipschitz constraints to ensure output stability and derive, for the first time, a theoretical worst-case coverage bound for standard CP under arbitrary attack magnitudes. Experiments on medium- and large-scale benchmarks (e.g., ImageNet) show that Lip-RCP reduces robust prediction set size by up to 42% over state-of-the-art methods while accelerating inference by 3.8×. Crucially, it strictly guarantees both nominal coverage ≥90% and finite-sample robust coverage—without compromising statistical validity.

1 citationsRead paper

Using SRv6 to access Edge Applications in 5G Networks

Dec 08, 2023StudentWorkshop@CoNEXT

To address suboptimal data paths and challenges in policy-driven resource scheduling when user equipment accesses edge applications in 5G multi-access edge computing (MEC), this paper proposes an end-to-end programmable path orchestration method deeply integrated with Segment Routing over IPv6 (SRv6). It is the first to natively embed SRv6 into the 5G standalone (SA) core network–edge cloud collaborative architecture, synergizing network slicing, decentralized user plane function (UPF) deployment, and a policy-driven control-plane coordination mechanism. The approach enables dynamic, service- and operator-policy-aware path orchestration and millisecond-scale rerouting. Simulation results demonstrate a 37% reduction in end-to-end latency, while edge service accessibility and policy compliance rate improve to 99.2%.

1 citationsRead paper
Recent publications

Latest Papers