AI Research Group
Institute of Cognitive Science
Research Area "Artificial Intelligence"
Institut für Kognitionswissenschaft
Albrechtstraße 28, 49076 Osnabrück
Tel: +49-541 969.3380, Fax:+49-541 969.3381
The artificial intelligence group focuses primarily on the modelling of higher cognitive abilities, for example on (classical and non-classical) reasoning, memory, and learning. The methods used for modelling these abilities include logic-based, algebraic, and co-algebraic approaches, programming paradigms, as well as methods used in neuroscience. From a more abstract perspective these interests can be embedded into the following questions stressing the cognitive aspect:
- How can we theoretically explain and practically model creativity of humans?
- How can the gap between symbolic and sub-symbolic representations be closed in artificial systems?
- How can machines process information at a semantic level?
Besides these more or less theoretical questions the AI group works on practical realizations in running software applications.
Main Research Areas of the Artificial Intelligence group:
- Analogical Reasoning and Metaphors:Goals of this research area are to develop syntactic, semantic, and algorithmic approaches for models of analogical reasoning and analogical learning. Strongly connected with this line of research is the interpretation of metaphorical expressions.
- Ontologies: In this research area, the AI group tries to develop a uniform framework for coding ontological and syntactic knowledge. Ontologies – in other words hierarchically structured background knowledge – are useful for a variety of technical applications.
- Knowledge Management: The goal is to develop mapping tools for cooperative work and learning scenarios. This area is strongly connected to research in document management.
- Algebraic Methods and Logic in Artificial Intelligence: We are interested in applying algebraic methods like category theory, universal algebra, co-algebras etc., as well as non-classical logic accounts and logic programming techniques to reasoning problems in AI.
- Symbolic/sub-symbolic representations: This research area attempts to model first-order inferences with neural networks and to extract conceptual knowledge from trained connectionist systems. The overall goal is to bridge the gap between symbolic and sub-symbolic representations.
- Learning Environments:This research area includes ICALL systems (Intelligent Computer Assisted Language Learning systems) as well as the whole range of media-based learning environments.