Supporting Defeasible Argumentation Processes over Relational Databases

ARIEL DAGUSTINI; SANTIAGO FULLADOZA; SEBASTIÁN GOTTIFREDI; MARCELO FALAPPA; CARLOS IVÁN CHESÑEVAR; GUILLERMO SIMARI,

LECTURE NOTES IN COMPUTER SCIENCE

Springer

Lugar: Berlin; Año: 2012

0302-9743

This paper introduces Database Integration for Defeasible Logic Programming (DBI-DeLP), a framework that integrates Defeasible Argumentation with Relational Databases that may be updated by other external applications, allowing the execution of argumentation processes based on massive external sources of data. Nowadays Defeasible Argumentation Systems in general (and DeLP in particular) build arguments based on the context of a single, xed logical program. This program contains both the rules and the data used to build arguments. If we want to include new knowledge in the program, we have to encode it manually to make it explicit. Defeasible argumentation technologies are often used to develop expert systems, as they allow them to give not only rened answers but also explanations. Nevertheless, in a production environment, such systems often need to be fed with large amounts of data. In this paper we formalize an approach for the extraction of data from a large information repository (based on one or several relational databases) that will be used to support arguments. This repository may be shared among agents or systems with dierent purposes, making it part of the community knowledge. In our scenario, we assume that the databases involved are updated by external independent applications, so that new data are made available to the expert system constantly.