click here, if navigation frame does not appear on the left (or more than once)


Michael Biehl
Professor in Computer Science,
Machine Learning and Computational Intelligence

University of Groningen
Bernoulli Institute for Mathematics,
Computer Science and Artificial Intelligence
Intelligent Systems Group
postal:   P.O. Box 407, 9700 AK Groningen, NL
visitors: Nijenborgh 9, 9747 AG Groningen, NL
              Room # 584, Bernoulliborg
Tel:        (+31) (0)50 363 3997    (Fax: ... 3800)                                                            
e-mail (official):     m dot biehl at rug dot nl
e-mail (private):    meikelbiehl at gmail dot com


Please note:
Should any open positions (internships, PhD, PostDoc etc.) become available in my working group, I will announce them here and through the usual channels.
I will not reply to unsolicited requests individually.


News and events:

April 2019: Special Session at ESANN 2019
Statistical Physics of Learning and Inference (click for details)
Organizers: M. Biehl, N. Caticha, M. Opper, T. Villmann

April 2019: Supervised Learning - An Introduction
Lectures given at the 30th Canary Islands Winter School on Astrophyiscs, Nov. 2018
Notes, slides and video lectures available from the e-book XXX Winter School
Lecture notes also published in Machine Learning Reports, Report number 01/2019
and available directly here.

April 2019: Mandy Lange-Geisler defended her PhD thesis
Hebbian learning approaches based on general inner products
and disance measures in non-Euclidean spaces

Congratulations, Dr. Mandy! :-)


Recent journal publications

March 2019:
Learning vector quantization and relevances in complex coefficient space
M. Straat, M. Kaden, M. Gay, T. Villmann, A. Lampe, U. Seiffert, M. Biehl, F. Melchert
Neural Computing and Applications, online, open access (2019).

February 2019:
Galaxy classification: A machine learning analysis of GAMA catalogue data
preprint available as arxiv e-print 1903.07749
A. Nolte, L. Wang, M. Bilicki, B. Holwerda, M. Biehl.
Neurocomputing 342: 172-190 (2019), available online.

December 2018:
Effect estimate comparison between the prescription sequence symmetry analysis (PSSA) and parallel group study designs: A systematic review, D.L. Idema, Y. Wang, M. Biehl, P.L. Horvatovich, E. Hak
PLoS ONE 13(12): e0208389 (2018, open access)

October 2018:
Statistical Mechanics of On-Line Learning Under Concept Drift
M. Straat, F. Abadi, C. Göpfert, B. Hammer, M. Biehl.
Entropy 20(10): 775 (2018) available online (open access)


Available code (GMLVQ and variants):

Python code (scikit-learn compatible)
A collection of LVQ variants
including GMLVQ is avalailable now (thanks to our colleagues at CITEC, Bielefeld University!)
direct download (zip)    github (including scikit-learn)

A Jave plug-in for WEKA
is available from our colleagues at the CI Group Mittweida
- github repository

GMLVQ demo code in Matlab (TM):
A collection of no-nonsense code in Matlab (TM) is available for
Generalized Matrix Relevance LVQ (GMLVQ), see here.
A more complete and more sophisticated Matlab (TM) implementation of
GMLVQ and several variants (due to K. Bunte) is available here.


Objective measures of scientific excellence :-) (as of June 2019)
 5936: citations in scholar.google,   (i-10 index: 109)
 2800: citations in web-of-science / publon (147 articles in WoS, 19.1 cit. per article)
 91: May index (May-factor) 2019/20 (May 2019)
 36: Hirsch index (h-index, h-factor) according to scholar.google
 25: Hirsch index (h-index, h-factor) according to web-of-science
 21: Publications among the 10% (or better) most frequently cited
         of their year and discipline (2007-2017 only) according to Web of Science (Jan. 2018)
 5: seminars co-organized at MPI für Physik komplexer Systeme, Dresden
 4: seminars co-organized at Schloss Dagstuhl (most recent one in June 2016)
 3: Erdös number (pretty much constant, recently)

For slightly more detailed information :-) check out my publications (navigation bar on the left)
Many "services" provide unauthorized and incorrect research or publication profiles. I occasionally check the ones listed in the menu bar (left), but for (almost) complete lists of my publications, citation analyses, frequently cited papers etc. I strongly recommend my Scholar Google profile or the Web-of-Science based Researcher ID:


Research topics:         (see publications for details)

Theory, algorithm development, and applications of machine learning
· computational intelligence, supervised learning (classification, regression), unsupervised learning (e.g. clustering)
· machine learning in, e.g., medical data analysis, tumor classification, bioinformatics, astronomy
· non-standard distance measures, adaptive metrics in supervised learning, feature selection and relevance learning
· low-dimensional representation and discriminative visualization of labeled data sets
· mathematical analysis, design and optimization of training algorithms in artificial intelligence

An overview of my earlier contributions to the field of machine learning


Scientific Computing, Modelling and Simulation
· equilibrium and off-equilibrium Kinetic Monte Carlo (KMC) simulations
· theory and modelling of crystal growth, thin films and surfaces
· off-lattice simulation of hetero-epitaxial growth: dislocations, self-organized nano-structures, surface alloys

An overview of research in growth and surfaces