Skip to ContentSkip to Navigation
ResearchBernoulli InstituteDepartmentsResearch and projects


Multi-Agent Systems Group


The Vici project 'Cognitive systems in interaction:
Logical and computational models of higher-order social cognition'

Software agents are capable of recursion in their social reasoning: A thinks about what B knows, about what B thinks that A knows, and so on. However, humans lose track of such reasoning after only two or three levels. If software agents work together with human team-mates, they need to take into account the limits of their social cognition.

In the Vici project, we investigate children's development and adults' limitations in applying higher-order reasoning, using a close-knit combination of empirical research and formal modelling. Whereas first-order social cognition has been intensely investigated, higher-order social cognition is far less well-understood. This Vici project aims to apply improved understanding of higher-order social reasoning to design realistic logics, ready for implementation in systems supporting mixed human-computer teams.

[ Link to the project website: ]

The NWO Physical Sciences Free Competition project
'STRATMAS. Strategies in multi-agent systems: From implicit to implementable'

Modelling intelligent and rational interaction in multi-agent systems has been one of the main issues in Artificial Intelligence that gained momentum in the last decade of the past century. The 'agents' under discussion mainly referred autonomous actors like software programs, robots and even humans. Interaction among them can be cooperative as well as non-cooperative.

This is now merging into broader studies of formal models of society, where computer science meets decision theory, game theory, and social choice theory, for instance in the study of rational deliberation and decision making. One connection that has been made here is 'social software', an interesting perspective, but as yet without a complete theory. There is a number of detailed studies of agents' knowledge, beliefs, preferences, and also of their long-term powers for influencing the outcomes of games. But we do not have a good theory of what may be the most crucial ingredient here: the plans or strategies that information-processing agents have for achieving these goals. In other words, their know-how in addition to their know-that. This is the main topic of this research program.

[ link to the project website: ]

Designing and Understanding Forensic Bayesian Networks with Arguments and Scenarios

Recent miscarriages of justice have increased the interest from legal practice in scientifically founded ways of treating evidence. Forensic statistics can provide such foundations. However, because of the communication gap between forensic statisticians, crime investigators and lawyers, statistical evidence is easily misinterpreted in court, resulting in wrong decisions. Therefore, methods must be developed to support the communication between the parties involved. Since lawyers are more used to thinking in terms of arguments and scenarios, we propose to develop methods that support argumentation- and narrative-based communication about statistical evidence, building on AI models of argumentation and scenario construction.

[ more information: ]

VENI-project Developing dynamic epistemic logic guided by old paradoxes and new techniques

Information pervades so many aspects of our daily lives that the current age has been dubbed "the information age". There are many scientific theories about information. In the last decade dynamic epistemic logic has developed as one of the main formal systems that are used to analyse the dynamics of information. These recent developments have mostly taken place within computer science and artificial intelligence. Philosophers have yet to benefit from these developments in order to solve some well-known problems in philosophical logic. The overall aim of the project is to further develop dynamic epistemic logic and to let it join forces with other scientific theories about information in order to gain insight in information and its dynamics.

[ more information: ]

Laatst gewijzigd:31 mei 2018 16:40