My main interest in research is in automated reasoning and knowledge representation. Since many years my focus is on applying dedution systems to real world problems. One of my major projects to this end is the Loganswer/Ratiolog project which aims at a natural language open domain question answering system. Nowadays IBM calls the technique we are using "cognitive computing".
In cognitive science there is a lot of work to desrcibe and to model human reasoning. Recently we found out, that our 1st order theorem prover Hyper is able to process Deontic Logic, which can be used very naturally to describe norms and policies.
With my partner Ulrike Barthelmess we are working on a second book project on "Künstliche Intelligenz - von Bergson über Proust zu Nabokov".
Ratiolog and his predecessor Loganswer are finished projects about deep question answering based on Computer Linguistic and Automated Deduction. In Rationlog the main goal was to incorporate models for human rationality.
For more see:The RatioLog Project
This is an approach to model human reasoning as it is investigated in cognitive psychology by means of symbolic logic and automated deduction. A very typical example for this is the well researched Wason selection task. There are publication to this topic and there is workshop series:
In this book project we are aiming at a combination of cultural and literary treatments of knowledge and remembering with modern computer science and artificial intelligence work. An example is the depicted cookie, a Madelaine, about which Marcel Proust wrote, that the taste of it triggered the memories about a certain episode of his life.