Research

A short list of the areas I've worked on in my time at the Institute of Artificial Intelligence Faculty of Computer Science of the Technische Universität Dresden. Latest things come first. You can find the various publications coming from that in my publications page.

Intelligent Systems in E-Commerce

An industrial collaboration with [hybris GmbH](https://en.wikipedia.org/wiki/Hybris_(company)), Munich. The project applied some techniques of artificial intelligence to a generic web shop system. A selection of fuzzy logic learning algorithms are used to determine online the customers interests, and the product presentations is adapted accordingly. In the second stage of the project we investigated collaborative filtering techniques for generating product recommendations.

Fluent Calculus

The Fluent Calculus is a logical framework to formalize and solve planning problems. I made a number of contributions ranging from the specification of recursive plans over an extension of the underlying equational theory to an embedding of a fragment of the Fluent Calculus into propositional logic, such that the power of Binary Decision Diagrams can be used to find shortest plans efficiently. The latter is topic of my PhD-thesis.

Neural Networks and Logic Programs

In joint work with Yvonne McIntyre (formerly Kalinke) and Steffen Hölldobler we proved, that the semantics of acceptable first order logical programs can be approximated arbitrarily well by neural networks.

Feigenbaum constant

An interesting problem of numerical mathematics is the calculation of the so called feigenbaum constant delta=4.669201... that appears in many fractal structures. I improved the accuracy of previous results from 108 to 435 decimal places by creating a new functional equation for it and approximating the solution numerically using variants of gradient descent.