CASUS Institute Seminar, Prof. Dr. Thomas Villmann, University of Applied Sciences Mittweida (Germany) and Saxon Inst. f. Computational Intelligence and Machine Learning (SICIM)

The development of smart and interpretable machine learning approaches is required for many application domains particularly in engineering and automotive. Yet, many other application domains for AI would also benefit if complex non-interpretable models could be replaced by more sophisticated and self-explaining approaches.

In this presentation, Prof. Thomas Villmann will highlight a classification model based on vector quantization strategies. The basic model can be easily adapted to reflect user specific requirements. Moreover, theoretical justifications for robustness and convergence are given.

In addition, an exemplary application will be given to illustrate the abilities and the properties of the approach.