Partenaires
Rechercher
Navigation
Accueil du site > English > Research Topics > Topic 4 - Reasoning and Decision > ADRIA team > Research topics
Research topics
Knowledge representation
Contributions can be clustered in three main groups: the representation of uncertainty, the representation of preferences, and bipolar representations. Besides, a recent cooperation with the LILAC team in IRIT deals with the assessment of trust among agents.
Models of reasoning
The team has a continuing activity on multiple source information fusion , , reasoning in the presence of information that is both uncertain and defeasible , and the revision of beliefs upon arrival of new information (as opposed to the contextual querying of a knowledge base) . The relevance of truth-functional calculi for epistemic reasoning tasks has been discussed at length, especially partial logic and Belnap logic , and three-valued logics for rough sets. Efforts to extend (...)
Argumentation, Negotiation
An argument gives a reason for believing a statement, choosing an action, adopting a goal, etc. Hence, arguments may have different forms, different strengths and may even be more or less relevant to the standpoints they support. Argumentation is seen as a reasoning process in which arguments are built and evaluated in order to increase or decrease the acceptability of a given standpoint. It is an important component of autonomous agents’ reasoning, namely for handling inconsistency in (...)
Machine learning
Machine learning aims at developing algorithms for inducing synthetic models for describing data (i.e. finite sets of examples and counter-examples), and designing predictive models. The interest of the group in uncertainty and preference representation and commonsense reasoning has led it to investigate diverse issues in relation to machine learning: extension of standard machine learning frameworks to take into account uncertain and/or imprecise information , , , , , , machine learning (...)
Real-world planning, Supply Chain Management
The resolution of real-world planning problems is a very difficult process. To obtain efficient algorithms it is necessary to respect a very restrictive framework (instantaneous actions, finite world, static environment…) . We seek to enlarge this framework to solve real-world problems: we are working on algorithms that generate an optimal solution to problems in which each action has an associated strictly positive cost , , . We can detect indispensable sets, sets of actions one of which (...)
composition of the team
The team consists of L. Amgoud, J.-C. Buisson, C. Cayrol, Mr Cooper, D. Dubois, F. Dupin of Saint-Cyr, B. Fade, H. Fargier, H. Farreny, M.-C. Lagasquie, J. Mengin, H. Prade, P. Régnier, G. Richard, Mr Serrurier, C. Thierry. S Benferhat, Mr Cayrol, F. Evrard, S Konieczny, J. Lang, Y. Luvisutto, L. Sais were it. J.-F. Bonnefon, B. Gaume, and J. Lang has activities associated with the team. A. Lorente, D. Roncier, S Senac, B. Bos, K. Santrisse, F. Boué successively assisted the team the (...)