Mission statement

The research unit KERMIT (Knowledge-based Systems) adopts a holistic view on mathematical and computational modelling, acknowledging the needs of our modern information society with a particular focus on the applied biological sciences. It strives to keep a unique balance between theoretical developments and practical applications, a strategy that has proven particularly successful, regarding the output, visibility and recognition of the team. It plays a pioneering role by promoting existing as well as developing new methods in a broad range of disciplines shown below.

Methodological expertise

KERMIT started out in 2000 with two men and a dog as a team in fuzzy set theory and preference modelling, and quickly expanded into the fields of artificial/computational intelligence and operations research. Over the years, it has evolved into a holistic team covering the entire data-to-decision pipeline. It provides methodological expertise in the following areas:

  • image processing and computer vision (e.g. mathematical morphology, filters)
  • knowledge-based systems (e.g. bioportals, ontologies)
  • dynamical modelling (e.g. differential equations, cellular automata, individual-based models)
  • data science and machine learning (e.g. time series prediction, structured prediction, deep learning, big data)
  • management of imprecision and uncertainty (e.g. probability, possibility and fuzzy set theory; stochastic modelling, theory of copulas)
  • optimization and operations research (e.g. preference modelling, multi-criteria decision making)


Doctoral degree for Ward Quaghebeur

On January 11, 2023, Ward Quaghebeur successfully defended his Ph.D. thesis "Hybrid models of dynamical systems: neural differential equations, Shapley value analysis, and illustrations in water systems" and was awarded the title of Doctor of Bioscience Engineering: Mathematical Modelling. Ward was supervised by Ingmar Nopens, Bernard De Baets and Elena Torfs. His research was supported by the Research Foundation - Flanders.

PhD Ward

All KERMIT news


Most recent journal publications
Biblio logo(702) Valid prediction intervals for regression problems
N. Dewolf, B. De Baets and W. Waegeman
Biblio logo(701) Combining natural language processing and multidimensional classifiers to predict and correct CMMS metadata
A. Deloose, G. Gysels, B. De Baets and J. Verwaeren
(2023) COMPUTERS IN INDUSTRY. 145, 103830.
Biblio logo(700) A deep recursive multi-scale feature fusion network for image super-resolution
F. Liu, X. Yang and B. De Baets
All KERMIT publications