Niels Taatgen
Professor of Cognitive Modeling · Department of Artificial Intelligence · University of Groningen
I build computer models of how people think and learn. Right now my main focus is AI in education: using models of knowledge and skills to give every student exercises that are neither too easy nor too hard, and testing in real classrooms whether that leads to better results. Alongside this I continue my long-running work on how skills transfer between tasks, how we multitask, and how we perceive time, much of it based on the ACT-R cognitive architecture and on PRIMs, an extension I developed.
I studied computer science and psychology in Groningen and obtained my PhD there (cum laude) in 1999. Between 2003 and 2009 I was also affiliated with Carnegie Mellon University, working with John Anderson. In 2009 I returned to Groningen as a full professor and started the cognitive modeling group. From 2020 to 2025 I was Research Director of the Bernoulli Institute, and I am affiliated with CogniGron, where we work with materials scientists on cognitively grounded neuromorphic computing.
Research
Each topic has its own page, most with a short interactive explanation for readers who are not familiar with cognitive modeling. An overview brings the themes together.
AI in Education
Knowledge graphs learned from student data, and personalized practice that keeps every learner in the "just right" zone. Current focus.
Explore →PRIMs & transfer of skills
How a skill built from primitive steps explains why training on one task can help on another.
Explore →Multitasking
Threaded cognition: when does doing two things at once work, and when does it fail?
Explore →Time perception
An internal clock that gets slower as it ticks, and what that does to our sense of time.
Explore →News
- Dec 2026 Remon Nicolai joins the group as a PhD student, co-supervised with Hedderik van Rijn.
- 2025 New paper with the CogniGron group: Girão, Tiotto, Borst, Taatgen & Chicca, "Phasic attractors for flexible and adaptive working memory in spiking recurrent neural networks," Neuromorphic Computing and Engineering.
- 2025 New paper on language acquisition: Ji, van Rij & Taatgen, "Simulating procedural discovery in early language acquisition: Domain-general cognition with contextual learning," Journal of Child Language.
- 2024 Ram Eshwar Kaundinya joins the group (2024–2028) for a PhD on Cognitive Architectures for Touch-Based Applications, co-supervised with Jelmer Borst.
Teaching
I teach an introductory course in Artificial Intelligence and two Master's courses, Cognitive Modeling and AI in Education. I supervise Bachelor's and Master's theses on cognitive modeling, learning, multitasking and time perception. The PRIMs tutorial (Jupyter notebooks for pyprims) contains example models and exercises that are free to use in teaching; slides are in the original PRIMs tutorial.
Contact
- n.a.taatgen@rug.nl
- Phone
- +31 50 363 6435
- Address
-
Department of Artificial Intelligence
Bernoulli Institute, University of Groningen
Nijenborgh 9, 9747 AG Groningen
The Netherlands