The CRIL in short

présentation

Lens Computer Science Research Lab (CRIL UMR 8188) is a joined laboratory between Université d’Artois and CNRS, grouping together 50 members, including researchers, lecturers, PhD students, postdocs and administrative or technical staff.

Learn

Artificial Intelligence Research and Applications

mots clés du CRIL

News (RSS)

Seminar 

Société Altifort

Seminar Constraint Games

Arnaud Lallouet

Constraint Games are a recent framework proposed to model and solve static games where Constraint Programming is used to express players preferences. In a recent work, we rethink the solving technique in terms of constraint propagation by considering players preferences as global constraints. It yields not only a more elegant but also a more efficient framework. Our new complete solver is faster than previous state-of-the-art and is able to find all pure Nash equilibria for some problems with 200 players. We also show that performances can greatly be improved for graphical games, allowing some games with 2000 players to be solved.

Learn

Seminar Developing a parallel CP (or SAT) solver through the exploitation of strong propagation methods.

Kostas Stergiou (University of Western Macedonia)

As a result of the considerable recent advances in parallel constraint solving, a number of quite efficient parallel CP and SAT solvers have been developed. Parallel constraint solving techniques are roughly divided in search space splitting and portfolio-based ones. In this talk we describe a novel scheme for developing a parallel CP (or SAT) solver through the exploitation of strong propagation methods. This scheme is orthogonal to the two general categories of parallelization methods and it is presented as a search algorithm consisting of a main process, which is a typical CP (or SAT) solver, aided by a number of coworkers running in parallel.

Learn

Accepted papers at IJCAI'17

This year, there will be 8 papers from CRIL to be presented at IJCAI 2017 :



Acceptability Semantics for Weighted Argumentation Frameworks by Leila Amgoud, Jonathan Ben-Naim, Srdjan Vesic, Dragan Doder


Measuring the Intensity of Attacks in Argumentation Graphs with Shapley Value
by Leila AMGOUD, Jonathan Ben-Naim, Srdjan Vesic


Efficiently Enforcing Path Consistency on Qualitative Constraint Networks by use of Abstraction by Michael Sioutis, Jean-François Condotta


Constraint-Based Symmetry Detection in General Game Playing by Frederic Koriche, Sylvain Lagrue, Eric Piette, Sébastien Tabary


On Neighborhood Singleton Consistencies by Kostas Stergiou, Anastasia Paparrizou


A Recursive Shortcut for CEGAR: Application To The Modal Logic K Satisfiability Problem by Jean-Marie Lagniez, Daniel Le Berre, Tiago de Lima, Valentin Montmirail


Automatic Synthesis of Smart Table Constraints by Abstraction of Table Constraints by Baudouin Le Charlier, Minh Thanh Khong, Christophe Lecoutre, Yves Deville


An Improved Decision-DNNF Compiler by Pierre Marquis, Jean-Marie Lagniez


Learn

Proposition de thèse financée au CRIL sur l’argumentation abstraite

Utilisation de l’argumentation pour automatiser l’exploitation des
avis clients pour la vente en ligne et pour les consultations
citoyennes : modélisation des attaques et supports entre arguments

Le but principal de cette thèse est d’étudier la notion de support en
argumentation. La première tâche sera d’étudier les faiblesses des
approches actuelles. Il faudra en particulier identifier les résultats
fournis par ces approches sur des exemples issus de cas réels (avis de
clients, textes issus de journaux, plates-formes de débats en ligne,
etc.), et étudier similarités et différences. Ensuite, le but sera
d’étudier les principes qui devraient gouverner ces sémantiques
d’attaque/support, et de proposer des sémantiques adéquates. Les
sémantiques proposées seront alors évaluées théoriquement (en termes
de propriétés satisfaites, de leur complexité d’implémentation, etc.)
et pratiquement (par rapport à des cas réels). Il faudra enfin
compléter le cadre formel construit pour permettre de prendre en
compte des votes positifs ou négatifs sur les arguments et/ou sur les
relations (de support et d’attaque), ainsi que d’autres informations
(réputation des utilisateurs, détection de trolls, etc.).

Pour plus d’information sur ce sujet, se référer à sa description complète.

Les candidats intéressés sont invités à contacter très rapidement
les encadrants de cette thèse:


Sébastien Konieczny (konieczny@cril.fr)
Pierre Marquis (marquis@cril.fr)
Srdjan Vesic (vesic@cril.fr)


Learn