DECIDE

Research directions

The DECIDE team intends to provide decision support solutions for decision makers facing heterogeneous and complex data. These data - text, signals, images, sensors streams, social networks interactions, decision making contexts, preferences of the decision makers, spatial, or even previous decisions - are the starting point of DECIDES's research activities.

Key informations

• Team leader : Romain Billot (Professor, IMT Atlantique),
• 6 full professors,
• 1 emeritus professor,
• 11 associate professors,
• 2 research engineers,
• 2 engineers,
• 1 post-doctoral researcher,
• 12 PhD student,
• Locations : the DECIDE team is located over 4 different institutions: ENSTA-Bretagne, IMT Atlantique, UBO, UBS.

Challenges

In order to facilitate the decision making act, DECIDE's ambition is to propose the decision makers solutions that allow them to:

• Identify, model, and understand the information extracted from data.
• Make reliable and robust decisions based on the extracted information.
• Justify decision recommendations to demonstrate their quality.
• Provide coherent recommendations according to the decision makers' needs.
• Model the different stages of the decision support process to guarantee that recommendations are both readable and traceable.
• Identify potentially uncertain preferences.
• Reach a compromise whenever decision makers have conflicting preferences.

Team structure

To tackle these scientific challenges, the DECIDE team works along 3 research axis : data, decision and information. The figure below summarizes this structure. The techniques used through the 3 axis are mainly data mining, machine learning, graph theory, optimization, data fusion. The decision makers are central to the activites of DECIDE, as it is shown on the figure.

Data

The "data" axis deals with the extraction of knowledge from large quantities of data (either structured, or partially structured) via automatic or semi-automatic methods.

This axis currently includes the following research areas : data mining (unsupervised and supervised learning), text mining, graph mining, and data visualisation.

Information

The "information" axis focuses on the agregation and the quality of data and information originating from heterogeneous sources.

More specifically, the team is interested in various measures of the quality of this aggregated information, in the relevance of the descriptors of visual content, in the quality of the information in accordance with the usage, and in the caracterization and the generalization of algorithmic properties of quality measures.

Decision

The "decision" axis aims to help one or more individuals, confronted with complex decision alternatives, possibly described by multiple conflicting consequences, to make decisions by providing them appropriate support tools.

This axis currently includes the following research areas: modelling and the elicitation of preferences, multi-criteria decision aiding, optimisation, mathematical programming, heuristics, metaheuristics and matheuristics.