26. – 30. August 2007, Dagstuhl Seminar 07351
Formal Models of Belief Change in Rational Agents
Giacomo Bonanno (University of California – Davis, US)
James P. Delgrande (Simon Fraser University – Burnaby, CA)
Jérôme Lang (Paul Sabatier University – Toulouse, FR)
Hans Rott (Universität Regensburg, DE)
Auskunft zu diesem Dagstuhl Seminar erteilt
Flexible Intelligenz statt starrer Logik. Bericht von Peter Zschunke, Associated Press, 04.09.2007 (German only)
Können Computer ihre Meinung ändern?
28.08.07 (German only)
The theory of belief revision studies how a rational agent should change its beliefs when receiving or perceiving new information about the environment. This new information could include objective properties of the actual world, occurrences of events, and, in the case of multiple agents, public or private communications among agents (possibly concerning their beliefs and preferences) as well as actions taken by other agents. Not surprisingly, this area has been of interest to researchers in different communities.
The initial research in belief change came from the philosophical community, wherein belief change was studied generally from a normative point of view (that is, providing axiomatic foundations about how rational agents should behave with respect to the information flux). Subsequently, computer scientists, especially in the artificial intelligence (AI) and the database (DB) communities, have been building on these results. Belief change, as studied by computer scientists, not only pays attention to behavioral properties characterizing evolving databases or knowledge bases, but must also address computational issues such as how to represent beliefs states in a concise way and how to efficiently compute the revision of a belief state.
The most important question in Game Theory is how to rationally form a belief about other players’ behavior and how to rationally revise those beliefs in light of observed actions. Traditionally Game Theory has relied mostly on probabilistic models of beliefs, although recent research has focused on qualitative aspects of belief change. A new branch of logic, called Dynamic Epistemic Logic, has emerged that investigates the epistemic foundations of game theory from the point of view of formal logic. Another, related, new field of research, called Social Software, maintains that mathematical models developed to reason about the knowledge and beliefs of a group of agents can be used to deepen our understanding of social interaction and aid in the design of successful social institutions. Social Software is the formal study of social procedures focusing on three aspects: (1) the logical and algorithmic structure of social procedures (the main contributors to this area are computer scientists), (2) knowledge and information (the main contributors to this area are logicians and philosophers), and (3) incentives (the main contributors are game theorists and economists).
There are various newly emerging links between the research areas mentioned above. The purpose of the Workshop was to bring together researches from all these different areas; these researchers normally do not meet together. Workshops such as this one promote an exchange of ideas and cross-fertilization across different fields.
We found the Workshop successful, especially on the following two achievements: first, the seminar made participants aware of a commonality of interests across different disciplines; second, it suggested new directions for research that will probably be taken up by researchers in the next couple of years.
Where is the field going? We can mention at least two emerging issues:
- the field is broadening with respect to theoretical underpinnings and is beginning to incorporate notions from game theory and social choice theory. It is also broadening with respect to application areas, moving beyond traditional areas in AI and database systems, to include areas in description logics, the semantic web and economics.
- there is an emerging focus on epistemic notions having to do with communicating, negotiating, competing, and collaborating agents. Dynamic epistemic logic seems to have an important role to play here.
Moreover, it looks like belief merging and iterated belief revision are still hot topics and will remain so for the next few years.
For the future, we plan further workshops to encourage continued interdisciplinary interactions.
Dagstuhl Seminar Series
- 09351: "Information processing, rational belief change and social interaction" (2009)
- 05321: "Belief Change in Rational Agents: Perspectives from Artificial Intelligence, Philosophy, and Economics" (2005)
- Artificial Intelligence / Robotics
- Interdisciplinary With Non-informatics-topic: Economics
- Decision Theory
- Belief change
- Rational agents
- Information economy
- Information processing