Configurations for Inference from Causal Statements: Preliminary Report

Philippe Besnard 1 Marie-Odile Cordier 1 Yves Moinard 2
1 DREAM - Diagnosing, Recommending Actions and Modelling
Inria Rennes – Bretagne Atlantique , IRISA-D7 - GESTION DES DONNÉES ET DE LA CONNAISSANCE
2 AIDA - Modeling and Machine Learning for Data Interpretation and Decision Assistance
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, INRIA Rennes
Abstract : When dealing with a cause (e.g., looking for something which could explain certain facts), cases involving some effect due to that cause are precious as such cases contribute to what the cause is. They must be reasoned upon if inference about causes is to take place. It thus seems like a good logic for causes would arise from a semantics based on collections of cases, to be called configurations that gather instances of a given cause yielding some effect(s). Such a view is in line with the famous counterfactual analysis of causation which provides the motivation for the logic presented here. Two crucial features of the counterfactual analysis of causation are transitivity, which is endorsed here, and the event-based formulation, which is given up here in favor of a fact-based approach. A reason is that the logic proposed is ultimately meant to deal with both deduction (given a cause, what is to hold?) and abduction (given the facts, what could be the cause?) thus paving the way to the inference of explanations. The logic developed is shown to enjoy many desirable traits. These traits form a basic kernel which can be modified but which cannot be extended significantly without losing the adequacy with the nature of causation rules.
Type de document :
Communication dans un congrès
Stefania Bandini and Sara Manzoni. AI*IA 2005 (9th Congress of the Italian Association for Artificial Intelligence), 2005, Milan, Italy. Springer, pp.282-285, 2005, LNAI
Liste complète des métadonnées

https://hal.inria.fr/inria-00511108
Contributeur : René Quiniou <>
Soumis le : lundi 23 août 2010 - 17:37:42
Dernière modification le : mercredi 16 mai 2018 - 11:23:02

Identifiants

  • HAL Id : inria-00511108, version 1

Citation

Philippe Besnard, Marie-Odile Cordier, Yves Moinard. Configurations for Inference from Causal Statements: Preliminary Report. Stefania Bandini and Sara Manzoni. AI*IA 2005 (9th Congress of the Italian Association for Artificial Intelligence), 2005, Milan, Italy. Springer, pp.282-285, 2005, LNAI. 〈inria-00511108〉

Partager

Métriques

Consultations de la notice

271