The Dynamic and Distributed Information Systems Group at the University of Zurich investigates the foundations and technical means for building systems that dynamically adapt to changes in the environment or in their goals and/or require the collaboration of many human or computational actors to achieve their goals. The implicit underlying hypothesis of these investigations are that dynamic and distributed systems need some means to reason about their tasks and to learn from their actions.
In order to reason systems need to be able to (i) collect some kind of information and (ii) draw conclusions from them. Whilst we make no assumptions about the nature of the information collected we assume that they usually come in either a typed graph or in some kind of probabilistic/statistcial format. We operationalize the former usually in the form of Semantic Web knowledge bases and the latter as probabilistic statements. These formats allow for the integration of widely distributed information sources such as the web. To learn from their action systems must contain some kind of machine learning component.
As a consequence most of our research projects either have a Semantic Web or a machine learning component – or both.
Currently active projects
- Analytic Fraud Detection
- Automatic Interpretation of Experimental NMR Data
- Cultural Adaptivity of User Interaction
e-Lico aims to assist users in Data Mining processes by suggesting them which steps to execute for a given data mining task.
The Localina project aims to develop a next-generation recommender system based on Semantic Web technologies and data mining techniques.
MetAgora combines quantitative and qualitative research methods to investigate the influencing factors in the development of on-line communities.
- Scalable Triple Stores
- The Smart Soft Ware House
- Talking to the Semantic Web
- Machine Learning
- Semantic Web
- MIT Process Handbook