Theme 4: Neural Computation and Neurotechnology

Foundations of Natural and Stochastic Computing

In our group we study the foundations of natural and stochastic computing. The group's research is centred around two research lines: Foundations and applications of probabilistic graphical models and Foundations of neuromorphic computing.

In the first line we study topics such as computational complexity, approximate inference algorithms, Bayesian statistics, non-parametric Bayes, explainable AI (in particular in the healthcare domain), and computational modelling of cognition, focused on the Predictive Processing account in neuroscience.

In the second line we study topics such as computational complexity, neuromorphic/hybrid algorithm design (in particular with respect to Green ICT), hardware-software co-design in neuromorphic engineering, brain-inspired models of computation, and philosophy of neuromorphic computation.

Our primary research methods are computational and formal modelling, conceptual analysis, algorithm design, and mathematical analysis of information representation and processing in frameworks such as Bayesian networks and spiking neural networks.

In our research we value explanations that are both mathematically rigorous and biologically plausible; that scale up to 'real world settings' embodied and embedded in a realistic environment, and that 'advance the theory' as well as 'explain the data'. We are committed to fostering an open, safe, and inclusive research environment for our group members.

Contact
Name:	Johan Kwisthout
Telephone:	024-3655977
Email:	j.kwisthout@donders.ru.nl
Visiting address:	Donders Centre for Cognition Thomas van Aquinostraat 4 6525 GD Nijmegen The Netherlands
Postal address:	Donders Centre for Cognition P.O. Box 9104 6500 HE Nijmegen The Netherlands

Key grants and prizes

DCC grant, "Turing's Razor: Computational realistic architectures for a Bayesian brain How to grow an internal model: A toolbox for the computational modeller" (220k).

Intel Neuromorphic Research Community grant “Neuromorphic algorithms and complexity” (21k).

DCC grant, "Understanding predictive processing in development: Modelling the generation of generative models" (212k).

NWO EW TOP grant, "Parameterized complexity of approximate Bayesian inferences" (217k).

DCC grant, "Turing's Razor: Computational realistic architectures for a Bayesian brain" (200k).

Key publications

Kwisthout, J. (2018). Approximate inference in Bayesian networks: Parameterized complexity results. International Journal of Approximate Reasoning, 93, 119-131.

Kwisthout, J., Bekkering, H., & Van Rooij, I. (2017). To be precise, the details don't matter: On predictive processing, precision, and level of detail of predictions. Brain and Cognition, 112, 84-91.

Hashkes-Pink, S., Van Rooij, I., & Kwisthout, J. (2017). Perception is in the details: A predictive coding account of the psychedelic phenomenon. Proceedings of CogSci'17.

Kwisthout, J. (2015). Most Frugal Explanations in Bayesian Networks. Artificial Intelligence, 218, 56-73.

Otoworowska, M., Riemens, J., Kamphuis, C., Wolfert, P., Vuurpijl, L., & Kwisthout, J. (2015). The Robo-havioral Methodology: Developing Neuroscience Theories with FOES. Proceedings of BNAIC'15.

Kwisthout, J., & Van Rooij, I. (2019). Computational resource demands of a predictive Bayesian brain. Computational Brain and Behaviour, 3(2), 174 - 18.

Links

Back to:
Theme 4:
Natural Computing & Neurotechnology

Research Group
Foundations of Natural and Stochastic Computing

Principal Investigator
Dr. Johan Kwisthout

Group members

Assistant Professor
Dr Mahyar Shahsavari

Postdoc
Dr Leila Bagheriye

PhDs
Arne Diehl
Nils Donselaar
Andi Lin (Jan 2023)
Erwin de Wolff