October 29, 2019
As one of two contributions, the ML2R-paper »Prototypes within Minimum Enclosing Balls« was selected for the Best Paper Award at this year’s International Conference on Artificial Neural Networks (ICANN 2019).
Prof. Dr. Christian Bauckhage, Lead Scientist »Machine Learning« at the Fraunhofer Institute for Intelligent Analysis and Information Systems IAIS and Professor for Computer Science at the University of Bonn, Rafet Sifa, Data Scientist at the Fraunhofer IAIS, and Dr. Tiansi Dong from the Bonn-Aachen International Center for Information Technology (b-it) worked together on the project as part of the »Kompetenzzentrum Maschinelles Lernen Rhein-Ruhr ML2R«.
In their paper, they present a method that filters out those characteristic data points from a given set of data that are particularly easy for human users to interpret (»explainable latent factors«). They were able to show that this approach can be implemented in the form of recurrent neural networks. »It is great fun to work together with excellent scientists across institutions and thus jointly advance top-level research on artificial intelligence. The BMBF-funded ML2R Competence Centre makes a major contribution to this«, said Bauckhage.
The scientific work has been published by Springer-Wissenschaftsverlag.
ICANN, rich in tradition, is regarded as one of the most important events in the field of machine learning and artificial intelligence. It is hosted by the European Neural Network Society ENNS and took place for the 28th time this year.