Explaining Argumentation over Alignment Agreements

Rouaa Wannous (1) , Cassia Trojahn dos Santos (2, 3)

1 L3I - Laboratoire Informatique, Image et Interaction - EA 2118
2 IRIT-MELODI - MEthodes et ingénierie des Langues, des Ontologies et du DIscours
3 UT2J - Université Toulouse - Jean Jaurès

Argumentation frameworks have been used as tools for reconciliating ontology alignments, through a series of proposals and counter-proposals, i.e., arguments. However, argumentation outcomes may not be so obvious to human users. Explaining the reasoning behind the argumentation process may help users to understand its outcome, and influence the user's confidence and acceptance on the results. This paper presents a mechanism for providing explanations on the way agreed alignments are established. Our mechanism is based on tracing each step of the argumentation process. These traces are then interpreted using a set of association rules, built from a decision tree that represents all possible statuses of arguments. From these rules, a multi-level explanation, in natural language, is provided to the users.

Mots clés: Argumentation; Ontology matching; Explanation

Domaines: Intelligence artificielle [cs.AI]; Apprentissage [cs.LG]; Logique en informatique [cs.LO]; Informatique et langage [cs.CL]

Rouaa Wannous, Cassia Trojahn dos Santos. Explaining Argumentation over Alignment Agreements. IEEE/WIC/ACM International Joint Conferences on Web Intelligence WI and Intelligent Agent Technologies IAT 2013, Nov 2013, Atlanta, United States. pp. 80-85. ⟨hal-01151026⟩