Welcome
The international workshop on intelligent systems and
computational intelligence is an annual event where we
provide a forum for discussing the state of the art,
ongoing work, open questions, project ideas, and
anything else related to the field, all in the friendly
atmosphere of the city of Groningen. The event itself is
free of charge. WISCI 2026 is planned for March 30th.
Join now!
00
00
00
00
Next planned workshop:
DAYS
HOURS
MINUTES
SECONDS
2026
03
30
Save the Date!
The Venue
Academy Building
Broerstraat 5 Groningen Room A7 The Academy Building is located in the city centre, 950 meters from the Central Station (13 minutes walk). Open on Google Maps The WISCI 2026 is organised by: Mariya Schumska, Roland Veen and Michael Biehl. email: team@wisci.euPast WISCI
Here you will find links to the programs of past WISCI editions. • WISCI 2025 • WISCI 2024 • WISCI 2023Participate
We invite you to partake in our workshop as a speaker or an interested audience. Give a talk about your research. Get inspired, ask questions, discuss the state of the art, and make new connections. For practical reasons we ask you to let us know if you will attend the workshop. The button below will take you to our short Google form.Program
Tentative, updated 12-03-2026Morning Session
09:00 Arrival and Welcome 09:20 Opening 09:30 Thomas Martinetz, Universität Lübeck, Germany Why good solutions dominate in highly over-parameterized neural networks 10:00 Michael Biehl, University of Groningen, The Netherlands The statistical physics of learning in over- and under-parameterized neural networks 10:30 Coffee break I 11:00 Michiel Straat, Bielefeld Universität, Germany Toward fluid flow control with differentiable simulators and data driven surrogates 11:20 Roland Veen, University of Groningen / MRC LMS, The Netherlands / UK Autoencoders versus PCA for feature extraction in FDG PET scans in neurodegenerative diseases 11:40 Sofie Lövdal, University of Groningen / University Medical Center Groningen, The Netherlands IRMA - recursive application of GMLVQ for increased classification performance 12:00 Lunch breakAfternoon Session
13:30 Thomas Villmann, Hochschule Mittweida, Germany Challenges for Prototype-based Reliable Classification Learning 14:00 Thomas Davies, Hochschule Mittweida, Germany Diminishing Returns - Integer Quantization and Effects on GLVQ Training 14:20 Lukas Hans, University of Groningen / TH Köln, The Netherlands / Germany Industrial Demand Forecasting with Scalable and Adaptive Approaches 14:50 Coffee break II 15:20 Lukas Bader, TH Köln / August Rüggeberg, Germany Predicting Customer Behavior in the Tool Industry Using MGLVQ 15:50 Jelle Dalenberg, University Medical Center Groningen, Objectherkenning.com, The Netherlands Solving Real-World Object Detection Problems at Objectherkenning.com 16:10 Open Discussion Anyone present is invited to take the floor. Brief presentations/pitches of project ideas, open problems, negative results, etc. or discussions of trends in machine learning, blended vs. single malt, ChatGPT / DeepThink / Manus / Mistral Le Chat, Quantum or Neuromorphic Computing, etc. 16:55 Closing
Welcome
The international workshop on
intelligent systems and computational
intelligence is an annual event where we
provide a forum for discussing the state
of the art, ongoing work, open
questions, project ideas, and anything
else related to the field, all in the
friendly atmosphere of the city of
Groningen. The event itself is free of
charge. WISCI 2026 is planned for March
30th. Join now!
00
00
00
00
DAYS
HOURS
MINUTES
SECONDS
Next planned
workshop:
2026
03
30
Save the Date!
The Venue
Academy Building
Broerstraat 5 Groningen Room A7 The Academy Building is located in the city centre, 950 meters from the Central Station (13 minutes walk). Open on Google Maps The WISCI 2026 is organised by: Mariya Schumska, Roland Veen and Michael Biehl. email: team@wisci.euPast WISCI
Here you will find links to the programs of past WISCI editions. • WISCI 2025 • WISCI 2024 • WISCI 2023Participate
We invite you to partake in our workshop as a speaker or an interested audience. Give a talk about your research. Get inspired, ask questions, discuss the state of the art, and make new connections. For practical reasons we ask you to let us know if you will attend the workshop. The button below will take you to our short Google form.Program
Tentative, updated 12-03-2026Morning Session
09:00 Arrival and Welcome 09:20 Opening 09:30 Thomas Martinetz, Universität Lübeck, Germany Why good solutions dominate in highly over-parameterized neural networks 10:00 Michael Biehl, University of Groningen, The Netherlands The statistical physics of learning in over- and under-parameterized neural networks 10:30 Coffee break I 11:00 Michiel Straat, Bielefeld Universität, Germany Toward fluid flow control with differentiable simulators and data driven surrogates 11:20 Roland Veen, University of Groningen / MRC LMS, The Netherlands / UK Autoencoders versus PCA for feature extraction in FDG PET scans in neurodegenerative diseases 11:40 Sofie Lövdal, University of Groningen / University Medical Center Groningen, The Netherlands IRMA - recursive application of GMLVQ for increased classification performance 12:00 Lunch breakAfternoon Session
13:30 Thomas Villmann, Hochschule Mittweida, Germany Challenges for Prototype-based Reliable Classification Learning 14:00 Thomas Davies, Hochschule Mittweida, Germany Diminishing Returns - Integer Quantization and Effects on GLVQ Training 14:20 Lukas Hans, University of Groningen / TH Köln, The Netherlands / Germany Industrial Demand Forecasting with Scalable and Adaptive Approaches 14:50 Coffee break II 15:20 Lukas Bader, TH Köln / August Rüggeberg, Germany Predicting Customer Behavior in the Tool Industry Using MGLVQ 15:50 Jelle Dalenberg, University Medical Center Groningen, Objectherkenning.com, The Netherlands Solving Real-World Object Detection Problems at Objectherkenning.com 16:10 Open Discussion Anyone present is invited to take the floor. Brief presentations/pitches of project ideas, open problems, negative results, etc. or discussions of trends in machine learning, blended vs. single malt, ChatGPT / DeepThink / Manus / Mistral Le Chat, Quantum or Neuromorphic Computing, etc. 16:55 Closing
