Programme

Conférenciers invités

• Simon Parsons (University of Lincoln) : Using AI to make agriculture more sustainable
Food is important. All of us rely on it, every day, to survive. At the same time, the methods that have been used to grow our food over the last 80 years raise concerns about long-term sustainability. Modern agriculture relies heavily on chemical herbicides and pesticides which have side-effects that cause ecological damage, and intensive agricultural practices have degraded soils and led to costly erosion. More sustainable agriculture is urgently needed, and artificial intelligence (AI) can help achieve it. In this talk, I will look at ways that AI can help to make our food supply more sustainable. All the examples will be based on work at the University of Lincoln, some that I have been involved in, and some that is the work of my colleagues. Broadly speaking this work falls into three categories: modifying existing practices by making use of AI; identifying new practices that are only feasible because of AI; and efforts to increase the pipeline of AI practitioners who are engaged in making agriculture sustainable.
• Tristan Cazenave (LAMSADE, Université Paris-Dauphine) : Utiliser l'IA pour améliorer l'IA
La thématique de la réflexivité de l'IA a été évoquée depuis les débuts de l'IA par plusieurs chercheurs, notamment par Jacques Pitrat avec le bootstrapping de l'IA. Nous aborderons dans cet exposé divers aspects d'amélioration de l'IA par l'IA, en particulier le renforcement mutuel des algorithmes de recherche et des algorithmes d'apprentissage machine. Nous verrons que ces idées ont des applications très diverses qui vont des jeux à la découverte d'algorithmes.
• Laurent Perron (Google Research) - Invité commun RADIA : CP-SAT in OR-Tools
The CP-SAT solver is developed by the Operations Research team at Google and is part of the OR-Tools open-source optimization suite. It is an implementation of a purely integral Constraint Programming solver on top of a SAT solver using Lazy Clause Generation. It draws its inspiration from the chuffed solver, and from the CP 2013 plenary by Peter Stuckey on Lazy Clause Generation. The CP-SAT solver improves upon the chuffed solver in two main directions. First, it uses a simplex alongside the SAT engine. Second, it implements and relies upon a portfolio of diverse workers for its search part. The use of the simplex brings the obvious advantages of a linear relaxation on the linear part of the full model. It also started the integration of MIP technology into CP-SAT. This is a huge endeavour, as MIP solvers are mature and complex. It includes presolve (which was already a part of CP-SAT), dual reductions, specific branching rules, cuts, reduced cost fixing, and more advanced techniques. It also allows the tight integration of the research from the Scheduling on MIP community along with the most advanced scheduling algorithms. This has enabled breakthroughs in solving and proving hard scheduling instances of the Job-Shop problems and Resource Constraint Project Scheduling Problems. Using a portfolio of different workers makes it easier to try new ideas and to incorporate orthogonal techniques with little complication, except controlling the explosion of potential workers. These workers can be categorized along multiple criteria like finding primal solutions (either using complete solvers, Local Search or Large Neighborhood Search), improving dual bounds, trying to reduce the problem with the help of continuous probing. This diversity of behaviors has increased the robustness of the solver, while the continuous sharing of information between workers has produced massive speedups when running multiple workers in parallel. All in all, CP-SAT is a state-of-the-art solver, with unsurpassed performance in the Constraint Programming community, breakthrough results on Scheduling benchmarks (with the closure of many open problems), and competitive results with the best MIP solvers (on purely integral problems).

-- LUNDI 1er JUILLET --

Session 1 (9h30-10h30) : Logique propositionnelle

• Nadia Creignou, Raïda Ktari and Odile Papini. Effacement des croyances en logique propositionnelle
• Nicolas François and Jean Lieber. Appliquer la logique des variations propositionnelles à la représentation de règles d'adaptation pour le raisonnement à partir de cas

Session 2 (10h50-12h) : Analyse comportementale en planification

• Salomé Lepers, Vincent Thomas and Olivier Buffet. Un cadre pour la planification consciente d’un observateur sous observabilité partielle
• Arnaud Lequen. Learning Interpretable Behaviour Classifiers for PDDL Planning

Session 3 (14h50-16h00) : Choix social

• Quentin Elsaesser, Patricia Everaere and Sébastien Konieczny. Agrégation de Jugements avec une Fiabilité Variable Inconnue
• Karl Jochen Micheel and Anaëlle Wilczynski. Equité dans le problème d’allocation répétée de maisons

Session 4 (16h20-18h00) : IA explicable

• Hénoïk Willot, Khaled Belahcene and Sébastien Destercke. Explications et caractérisation de décisions équitables
• Léo Saulières, Martin C. Cooper and Florence Dupin de Saint-Cyr. Backward explanation via redefinition of predicates
• Maxime Alaarabiou, Nicolas Delestre and Laurent Vercouter. Explicabilité en Apprentissage par Renforcement : vers une Taxinomie Unifiée

-- MARDI 2 JUILLET --

Session 5 (10h20-12h00) : Argumentation

• Jérôme Delobelle, Jean-Guy Mailly and Julien Rossit. Raisonnement Approximatif pour l’Acceptabilité des Arguments en Argumentation Abstraite
• Marie-Christine Lagasquie-Schiex, Jean-Guy Mailly and Antonio Yuste-Ginel. Gestion des supports dans les systèmes d’argumentation incomplets
• Yann Munro, Isabelle Bloch, Mohamed Chetouani, Catherine Pelachaud and Marie-Jeanne Lesot. Sémantique agrégative graduelle pour les systèmes d'argumentation bipolaires pondérés

11h50-12h00 : Assemblée générale JIAF

Session 6 (14h50-16h00) : Raisonnement par analogie

• Yves Lepage and Miguel Couceiro. Analogie et moyenne généralisée
• Stergos Afantenos, Henri Prade, Gilles Richard and Leonardo Cortez-Bernardes. Proportions analogiques et créativité: une étude préliminaire

Session 7 (16h20-18h00) : Planification hiérarchique ou temporelle

• Maxence Grand, Damien Pellier and Humbert Fiorino. Apprentissage de domaines HDDL à partir d'observations partielles et bruitées
• Nicolas Cavrel, Humbert Fiorino and Damien Pellier. Extension de la planification HTN aux problèmes temporels
• Ajdin Sumic and Thierry Vidal. A more efficient and informed algorithm to check Weak Controllability of Simple Temporal Networks with Uncertainty

-- MERCREDI 3 JUILLET --

Session 8 (10h20-11h40) : Jeux et IA

• Junkang Li, Bruno Zanuttini and Véronique Ventos. Combinatorial Games with Incomplete Information
• Conférencier invité : Tristan Cazenave. Utiliser l'IA pour améliorer l'IA