The MRIM Team is one of leading research group in France in the areas of Information Retrieval and Information Access. The MRIM Team studies models and algorithms for efficient access to information contained in large, multimedia, multilingual and semi-structured data collections.

 

The activities of the team are organized around several main research topics :

 

• Modeling multimedia data in the context of Information Retrieval and Information Filtering. This topic covers the definition of adequate indexing units (ie concepts, structured representation, ...) and specific models and algorithms dedicated to image, audio and video retrieval.

 

• Managing textual data for multilingual documents using conceptual indexes. This topic covers the definition of adequate automatic natural language processing, the integration of linguistic ressources and the extraction of concepts in order to provide a conceptual intermedia representation of documents.

 

• Design of personalized systems for information filtering. The challenge here is to design new effective interaction modalities supporting long retrieval sessions between users and the system. We also work on the collaborative aspects between users in order to enhance the information filtering process.

 

• Machine learning models and methods for information access. We are exploring, in this research topic, machine learning models for information access (categorisation, clustering, retrieval). More specifically we address problems related to large scale multimedia categorisation, similarity learning, and data mining in social networks and multimedia, multilingual documents collections.

 

• Contextual mobile information access. This new research domain is dedicated to the development of specific embedded systems allowing users to access to information in a mobile context. The underlying research leads to specific human interfaces development and image processing algoritms in order to allow information access on mobile devices (ex : mobile phones)

 

• System experimentation and evaluation. We are actively involved in the evaluation of our systems and prototypes. This includes a continuous analysis of system quality, usablility studies, and a participation to standard IR evaluation campaigns (TREC, INEX, CLEF). We participate to specific challenges (VideOlympics, Star Challenge).

 

We also develop specific applications based on medical data, News data, scientific libraries, etc

 

Since September 2003, a new activity related to Interactions among Internet Applications has been launched. The emergence of Web services technology and standards has radically lowered the cost barriers for exposing, reusing and sharing software applications over the Web. Organisations have now new opportunities to conduct business by enabling new access channels to their enterprise information systems through Web services. This trend has led to a paradigm known as Service-Oriented Architectures (SOA) wherein information and computational resources are abstracted as software services which are then interconnected using standard Web technology.

 

With this setting the team MRIM aims at proposing models and high-level languages for efficient management of interactions between resources (applications and/or information) available on Internet (i.e. accessible via Web services). More precisely these models and languages will allow organisations to describe, localise and integrate existing services and to compose new ones.