Background and challenges

This context has stimulated the search for new methods allowing the joint analysis of several data sets (structured data, multi-block, multi-channel) in many fields, such as:

• Machine Learning, where several approaches are considered for the processing of multi-source data (matrix factorisation, probabilistic approach).
• Chemometrics, where different methods are proposed to establish a chemical mapping of samples using several analytical techniques (generalisations of canonical analysis, NIPALS algorithm and tensor decompositions)
• Bioinformatics, where integrative methodological approaches allow the most complete picture possible of the dynamics of molecular systems to be drawn. 

In order to contribute to meeting the challenge of analysing and exploiting these multi-source data from an exploratory, but also predictive perspective, it is essential to bring together different viewpoints, practices and paradigms in order to reconcile these different approaches. It is also necessary to encourage collaboration between "method generators" and "data generators" in the various application fields.

This is the challenge that the MIMS consortium proposes to take up, by bringing together an interdisciplinary community working on approaches to the analysis and integration of multi-source data.

Goals

MIMS is a multidisciplinary consortium gathering more than 60 researchers, whose objective is to examine the analysis and exploitation of multi-source data, both in an exploratory and predictive perspective.

This consortium brings together multidisciplinary skills: information processing, biological sciences and analytics. The implementation of this multi-disciplinarity and its management will be based on the sharing of data, practices and methods between the partners, with the aim of formalising a scientific project to meet a common challenge: the optimal analysis of multi-source data for exploratory and predictive purposes.

Contacts :

Units involved and partners

INRAE participants

Partners