Koji Hasebe
Department of Computer Science, University of Tsukuba
Wednesday, July 5th, 2023, 14:00
Amphi F107, INRIA Montbonnot
entrée libre
Abstract
Opinion Dynamics is a process of public opinion formation through interactions between agents in a group. Since the first model was proposed more than half a century ago, there have been numerous studies on opinion dynamics. Many of the existing models present each agent's opinion as a numerical value. However, in such a model, the basis for the opinions has been abstracted, although the opinons we hold are formed based on our knowledge about the environment and the values. The objective of this study is to elucidate the mechanism of public opinion formation based on the knowledge held by agents and its relation with the process of cultural knowledge evolution among agents. For this objective, we propose a model in which networked agents update their knowledge through interactions. Here, each agent's knowledge and values are described by an ontology, and based on them, the agent's opinion is represented as a numerical value.
This research is on-going in collaboration with Jérôme Euzenat. In this talk, we present initial ideas and early results of our study.