CENTAL Louvain-La-Neuve

Full title: Expertise Localization from Informal Sources & Information Technologies
Start: January 2011
Duration: 36 months
Funding: WIST 3.0 (Région Wallonne)

The knowledge creation, dissemination and sharing has been a focus of concern for organizations since the advent of knowledge society. They are especially experiencing two strategic problems. The former consists in determining which kinds of knowledge are crucial for their activity, while the latter is the iden- tification of their experts. Who are they ? What are their competencies and skills ?

The main difficulty of these questions lies in the time component : hot topics appear and disappear continuously and new expertises are created every day. This new ever-changing pace has invalidated the traditional approach consisting in manually reporting links between skills and people in grids. This approach is time consuming and costly. Besides, it is inflexible and generally the grids are not regularly updated.

Yet people are communicating on a daily basis about their competencies and skills, inside and outside the organization : feeding blogs, asking and answering questions in emails, taking part in meetings focused on specific topics...

This project aims specifically to automate the analysis of these informal sources in order to support the management (identification, mapping and search) of knowledge, skills and people (experts) in an organization. A collection of soft- ware components will be developped to tackle this challenge. In the conception of these tools, four topics are chiefly targeted :

  • Legal aspects of the solution. The nature of some information sources (e.g. emails) raises issues and concerns about the privacy rights. Therefore, it is essential to establish a legal framework within which the solution could be implemented.
  • Information extraction and its preprocessing.The extraction is carried out based on informal sources. This means applying techniques from natural language processing and text mining fields to retrieve the relevant material.
  • Documents, skills and people mapping. The organization of the extracted information in the form of a graph needs tools from data mining and graph theory to infer the links (and their weights) between people, skills and documents.
  • Search for experts. The development of a search engine to find experts in organizations requires the use of information retrieval based technologies.


  • Adrien Dessy (UCLouvain/CENTAL)
  • Ilkka Kivimäki (UCLouvain/ICTEAM)
  • Feyrouze Omrani (FUNDP/CRID)
  • Dries Verdegem (ULB/IRIDIA)


  • Prof. Hugues Bersini
  • Prof. Cédrick Fairon
  • Prof. François Fouss
  • Kevin Françoisse
  • Prof. Pascal Francq
  • Prof. Yves Poullet
  • Prof. Marco Saerens

Industrial partners