Short portrait Cognitive Systems -- More than 15 years experience in comprehensible machine learning
In the research domain Cognitive Systems (CogSys) we are concerned with the development of approaches, concepts, and methods for design, description, construction, and analysis of artificial intelligence systems based on cognitive principles. On the one hand, we use findings on cognitive processes as inspiration to create artificial systems ('psychonics'), on the other hand, we are interested in computational modeling of cognitive phenomena. Therefore, in our research we combine empirical studies, development of algorithms, and their testing in different areas of application. Main topics of our group are induction and learning as well as planning and problem solving in single- and multi-agent settings. We have a long research tradition in inductive programming, that is inductive synthesis of (recursive) functional or logic programs from incomplete specifications (e.g., input/output examples). Inductive programming is a powerful and general approach to learning productive rules from experience. Our research addresses comprehensible and explainable artificial intelligence. Here our focus is on symbolic/knowledge-level approaches in machine learning (white box learning). Furthermore, we investigate analogical reasoning as a powerful approach to problem solving and as a special mechanism of knowledge acquisition. Additionally, we develop methods to generate verbal and visual explanations for learned classifiers, especially black-box approaches for image classification. Current application areas are identification of irrelevant digital objects, diagnostics in industry 4.0, facial expression analysis and classification, cognitive models of concept learning, intelligent tutor systems for mathematics and programming.