Computational Modeling of Human Social Cognition

Last updated on May 24, 2026

Pedestrians in Paris. Data was used for our crowd modeling research: We developed a model based on Festinger’s Social Comparison Theory (Social Psychology).

Alongside the mainstream of our research work on social intelligence in robots, we also investigate computational models of human social cognition, which we successfully compare to data from humans. There are three principal areas in which we work:

Most recently, we have begun to explore modeling the human mirror-neuron system at a process level. The mirror-neuron system is tied to human ability to infer the intent of others (indeed, even detect that motion is intentional), understand their actions, and interact with them. We are working on a novel plan recognition approach called mirroring (see also our plan-recognition research) which allows agents to understand others in terms of their own knowledge.
- E. Bonchek-Dokow and G. A. Kaminka. Towards computational models of intention detection and intention prediction. Cognitive Systems Research, 28(1):44–79, 2014.
- M. Vered, G. A. Kaminka, and S. Biham. Online goal recognition through mirroring: Humans and agents. In Proceedings of the Annual Conference on Advances in Cognitive Systems, 2016.
- Treger, S. and G. A. Kaminka. Towards Computational Modeling of Human Goal Recognition. Frontiers in Artificial Intelligence, 4, 2022.
Modeling social comparison processes and their role in group cohesion. We have developed an algorithmic model of Festinger’s Social Comparison Theory, and used it to simulate human group behavior:
- N. Fridman and G. A. Kaminka. Modeling pedestrian crowd behavior based on a cognitive model of social comparison theory. Computational and Mathematical Organizational Theory, 16(4):348–372, 2010.
- N. Fridman and G. A. Kaminka. Towards a computational model of social comparison: Some implications for the cognitive architecture. Cognitive Systems Research, 12(2):186–197, 2011.
- G. A. Kaminka and N. Fridman. Simulating Urban Pedestrian Crowds of Different Cultures. ACM Transactions on Intelligent Systems and Technology, 9(3), 2018.
We used qualitative reasoning (QR) to model crowd-level behavior and evolutions of demonstrations, successfully predicting what demonstrations will turn violent.
- N. Fridman and G. A. Kaminka. Using qualitative reasoning for social simulation of crowds. ACM Transactions on Intelligent Systems and Technology, 4(3), 2013.