1. médialab Sciences Po
  2. Team
  3. Audrey Baneyx
Audrey Baneyx profile picture

Audrey Baneyx

I have a PhD in artificial intelligence from the Paris 6 university and I'm interested in XXXX

Research Engineer

  • Mail: audrey.baneyx●sciencespo.fr

My research have situated themselves at the confluences of artificial intelligence, in particular the modeling of knowledge, and automatic natural language processing. After working specifically for a few years on textual analysis and the representation of knowledge in medicine, notably at the Inserm, I have since four years reinvested my methodological knowledge and techniques in service of social sciences.

Within the médialab, I intervene as a project leader and expert in the modeling  knowledge. My research activities have three major axis :

1. Think of different paths to resolve the problem of heterogeneity of digital data with the help of ontologies in order to improve our understanding of knowledge communities. The research led concerned the construction of epistemic networks centered around the way that communities form, how they evolve, interact and, more generally, the style of collaboration process they produce. There is a double challenge: On the one hand, to characterize the emergence and entanglement of these communities at different levels of granularity and, on the other hand, exhibit the objects of collaborations between individuals within communities.

2. To make visible the complexity of scientific production in the wide sense, it's mechanisms and functioning (ScholarscapeSciencescape). At a time where universities are set in international rankings, the issues are important for a field like social sciences that must be able to make sense of it's activity in order to evaluate itself while the disciplines that compose it are badly represented in the wide bases of normalized bibliometrics data (Scopus, Web of Science for example).

3. To develop automatic natural language processing tools adapted to social science research that permits to go from the constitution of a digital textual corpus (ANTA v2) to the modeling of the knowledge that it contains.