Michael R. Berthold | |
|---|---|
 Michael presenting at the 2022 KNIME Fall Summit | |
| Nationality | German |
| Occupation(s) | Computer scientist, entrepreneur, and author |
| Awards | KS Fu Award, the North American Fuzzy Information Processing Society Fellow, The Institute of Electrical and Electronics Engineers (IEEE) Honorary Professor, Óbuda University, Budapest |
| Academic background | |
| Alma mater | University of Karlsruhe, Germany |
| Academic work | |
| Institutions | Konstanz University, Germany |
Michael R. Berthold is a German computer scientist, entrepreneur, academic and author. He is a professor, and chair for bioinformatics and information mining at Konstanz University, and an honorary professor at Óbuda University.[1] He is also the co-founder of KNIME, and is serving as a president and CEO of KNIME AG since 2017.[2]
Berthold has authored over 250 publications while focusing his research on usage of machine learning methods for the interactive analysis of large information repositories. He is the editor and co-author of textbooks, including, Guide To Intelligent Data Science, and Intelligent Data Analysis.[3]
Berthold is a Fellow of the Institute of Electrical and Electronics Engineers (IEEE), the past president of the North American Fuzzy Information Processing Society,[4] and past president of the IEEE Systems, Man, and Cybernetics Society. He is an associate editor of Data Mining and Knowledge Discovery (DMKD),[5] Knowledge and Information Systems (KIS), Journal of Cheminformatics (JCIS),[6] and International Journal of Computational Intelligence in Bioinformatics and Systems Biology (IJCIBSB). He has been involved in the organization of various conferences, most notably the IDA-series of symposia on Intelligent Data Analysis.[7]
Berthold was born in 1966 in Stuttgart, Germany. He received his MSc degree in computer science in 1992, and his Dr.rer.nat. degree in 1997, both from Karlsruhe University.[8]
Berthold started his academic career as a visiting researcher at Carnegie Mellon University in 1991. He then held appointments as a visiting researcher at the University of Sydney in 1994, and as a researcher at the University of Karlsruhe in 1993. From 1997 till 2000, he was a BISC Research Fellow and lecturer at the University of California, Berkeley. Since 2003, he is a full professor, and chair for bioinformatics and information mining at Konstanz University, Germany.[8] In 2017 he took a leave of absence to become full-time CEO at KNIME AG, Zurich, Switzerland.[9]
At IEEE, Berthold served as a president of the IEEE System, Man, and Cybernetics Society from 2010 till 2011.[10]
Berthold has focused his research on large, and heterogeneous data sources, with particular focus on methods from AI (rule learning, neural networks, fuzzy logic and general machine learning).[11]
Berthold has published on methods to extract fuzzy models from data based on constructive methods to build probabilistic neural networks.[12] He developed similar algorithms for the extraction of fuzzy rule models.[13] He then extended those models beyond classification and invented algorithms to extract regression models, so-called fuzzy graphs from data automatically.[13]
At Konstanz University, Berthold initiated a European project (EU FP7 BISON) that focused on bisociative methods to create insights from diverse data sources. The consortium created output summarized in the resulting edited volume Bisociative Knowledge Discovery.[14]
Berthold was the first to introduce the idea of widened machine learning which draws on parallel resources to improve model accuracy rather than the usual focus on speed-up. He discussed a number of generic ways of tuning data mining algorithms while providing a series of experiments.[15][16] Later on, he conducted an in-depth analysis of the concept of Widened Data Mining, which aims at reducing the impact of heuristics by exploring more than just one suitable solution at each step.[17]
In 2017, Berthold and his team proposed the bucket selector, a model-independent randomized selection strategy, with the capability to perform better than existing selection strategies in cases without a diversity measure.[18]
