Dr. Santiago Ontañón and colleagues publish article in Artificial Intelligence Journal


August 31, 2012 — “A Defeasible Reasoning Model of Inductive Concept Learning from Examples and Communication,” a research paper written by Dr. Santiago Ontañón, assistant professor of computer science, in collaboration with Lluís Godo, Enric Plaza and Pilar Dellunde, from the Artificial Intelligence Research Institute in Barcelona, Spain, was accepted for publication in the Artificial Intelligence Journal.

Inductive generalization is the basis for machine learning methods that learn general hypotheses from examples. However, there has been little effort towards logical models of inductive generalization. This lack of a formal logical model of induction has hindered the development of approaches that combine induction with other forms of logical reasoning. This theoretical paper represents a first step towards a logical model of inductive generalization, and the authors show that it can be seen as a rather well-behaved non-monotonic logic. As an illustration of the utility of the model, the authors demonstrate that it allows the integration of inductive generalization with another form of logical inference (namely computational argumentation), and allows for a sound logical model of multiagent learning.