Université de Rennes 1


Computer science is focused on processing more and more data in order to answer more and more sophisticated questions. To meet this goal, and get smart systems, it is also well-recognized that knowledge of the domain, of the user, of the context, is the key. The department "Data and knowledge management' contributes to this challenging issue. 

More precisely, the objectives are to propose new algorithms and data structures for efficiently storing, retrieving, visualizing the data;

- to define adequate knowledge representation languages, taking into account uncertainty and imprecision, by proposing adequate data types and reasoning modes; 

- to propose methods to go from data to knowledge, enabling automatic acquisition and discovery of interesting patterns from large and often transient data;

- to keep data and knowledge in agreement when both are evolving along time, leading to update the model knowledge to fit the newly discovered data, 

- or to act on the data producer (a system, a user…) to fit to the reference model; and finally to ensure the privacy of capitalized data and knowledge, point that is crucial at long term.

The tools for managing and exploiting data, acquiring knowledge must deal with a huge amount of information (e.g. genomic data), often dynamic (understanding and monitoring dynamic systems is a key task) and often uncertain/imprecise (sensor accuracy but also abstraction induced by modelling). The tools must be flexible enough and user-adapted, in order to help the user express his/her information needs in a natural way (for example by letting him/her specify preferences which involve vague terms from the natural language) and to provide him/her with answers which are as rich as possible.

The Irisa teams involved in the DKM department share interest and competency in these various aspects of data and knowledge management (complex data, uncertain data, large-scale data, evolutive knowledge, and flexible retrieval). The research group shares a common background and foundations as logic, constraint and logic programming, artificial intelligence techniques, learning and data mining symbolic tools.

The application domains are web search engines, intelligent transportation systems, geographical information systems, e-health, on-line diagnosis and repair of industrial services and of software systems; decision support systems for environmental issues; bioinformatics and genomics.

Partner Status: