This pages lists the MSc. Theses topics currently available in our group. Don't hesitate to contact the respective person if you are interested in one of the topics. If you would like to write a thesis about your own idea you can propose it to the person most related to what you plan to do or you can contact Prof. Bernstein directly.



Proofs that you can trust in

Automatic negotiation of contracts is a challenging task. It requires unambiguous and commonly agreed on semantics and trust. In the Semantic Web, the semantics is covered by the W3C standard OWL2. Two agents can agree on terms that have well-defined semantics. Yet, there do not exist mechanisms how one agent can know whether to trust the claims that her opponent makes.

In this Master's thesis you will investigate how trust can be incorporated in a multi-agent system for the Semantic Web. You will compare different approaches for authenticating proofs and evaluate their pros and cons. Finally, you will give a prototypical implementation for the best solutions to show that your findings also work in practice.

Contact: Thomas Scharrenbach


Learning Conflict-Patterns

Conflicts can cause severe problems in OWL2 knowledge bases. As OWL2 is the de-facto standard for expressing information in the Semantic Web, it is hence desirable to either repair, ignore or prevent conflicts. However, even discovering conflicts for expressive knowledge bases is a 2NEXPTIME complete problem.

In you thesis you will investigate whether and how patterns for conflicts can be learned. This includes developing a framework which allows for a meta-description of OWL2 axioms, as well as an investigation of suitable learning procedures. You will, finally, evaluate your results on real-world data to show the usability of your approach.

Contact: Thomas Scharrenbach



Create a mapping and retrieval algorithm for the Hilbert curve that works for a n-dimensional space. So far we have an algorithm that optimizes the retrieval from the hilbert curve and is used as an index structure for 2-dimensional objects. Your task is to extend and evaluate the algorithm to work with an arbitrary number of dimensions.
This will be especially helpful for spatial and / or spatio-temporal indexes which are used in geographical information systems (GIS).

Contact: Jonas Tappolet


Do you really know your place?

You will implement a game that can be used to get an idea of the meaning of peoples' conception of local natural language expressions like "near". Two persons have to decide independently from each other if two presented locations are near/ not near (close; near/ far; next/ not next) to each other. The quickest player gets a point (so that we will get intuitive answers). These two persons should know the region from which we present the locations very good (Switzerland). Pairs of locations will be provided by a database with swiss place names.

Contact: Iris Helming