Techniques for reducing the complexity of algorithms to solve problems with uncertainty and preferences
Knowledge compilation is a family of approaches for addressing the
intractability of a number of difficult problems (beyond NP). A problem instance is "compiled" (pre
processed, typically transformed into an instance of another language) in an off-line phase in order to support the pertinent requests in polytime.
The aim of the ANITI KC project is to draw compilation maps (roughly, comparative complexity analyses), and to develop and experiment compilation models and algorithms for the on-line optimization of problems dealing with preferences and/or uncertainties, be the
uncertainty/preference model quantitative (e.g. GAI nets, Bayesian nets, Temporal CSPs) or qualitative (e.g. CP nets, logical approaches, point and interval algebra). Our domain of application includes temporal planning problems, configuration problems, and game theory.
Chair holder : Hélène Fargier, CNRS
Co- chairs :
- Christian Artigues (CNRS, LAAS)
- Romain Guillaume (UT2J, IRIT)
- Jérôme Mengin (UT3, IRIT), C.dric Pralet (ONERA)
Senior collaborating researchers
- Florent Teichteil-Konigsbuch (Airbus),
- Guillaume Poveda (Airbus) – Forthcoming
- Louis Riviere, since July 2020
- Hieu Tran Tron, since January 2021 Forthcoming joint Ph. D with ISAE/EMAC (E. Vareille/P. Gaborit), not funded by ANITI
- Nicolas Schmidt, since February 2020
Computer-assisted choice, with Hélène Fargier
Curious, Hélène Fargier is interested in ideas and projects as much as she enjoys meeting people to compare her points of view. An open attitude that she also adopts in matters of artificial intelligence, in her role as chair holder.