Meet Vincent Vandewalle, a new 3IA Côte d'Azur chair holder!

Published on September 12, 2022 Updated on October 24, 2022

Could you introduce yourself?

I followed the AgroParisTech, an engineering school in life sciences, and I quickly became fascinated by mathematical modeling. This is why I specialized in probability and statistics by following the M2 d'Orsay co-accredited by the school. I continued with a thesis in statistics on semi-supervised learning at the University of Lille. From 2010 to 2022, I was an Assistant Professor in the STID department (Statistics and Business Intelligence) of the IUT of the University of Lille. I was a member of two research teams: the ULR METRICS on the evaluation of health technologies and medical practices, on the other hand, the Inria MODAL team on the design and study of generative models for learning. I have focused my research activities on statistical learning and more precisely on clustering which aims at organizing objects into a finite number of groups based on their characteristics. My favorite topic is model-based clustering, it allows by modeling the data generation process to cluster potentially any kind of data. A part of my research has been directly motivated by applications such as medicine. I am also interested in the dissemination of my research as packages. Most of my teaching activities were in the DUT STID (now BUT STID), a post-bac formation focused very early on some core elements of data science (databases, programming, modeling). I had been head of the STID department of Roubaix, and particularly involved in updating the contents of the training, the development of interdisciplinarity between computer science and statistics, the development of work-study programs, and contacts with companies. Since September 2022, I have the great pleasure to be a Professor at the University Côte d’Azur and member of the Laboratoire J.A. Dieudonné and of the Inria MAASAI team.

Could you explain the topic of your chair?

Data arrive at an increasing rate, described by an increasing number of variables that are more and more diverse, so it becomes impossible for a human to directly derive usable information from theses data. Learning models then play the role of a constantly updated filter between the data and the human. My chair is about finding structures in these heterogeneous data. This search in based on generative models which are able to find latent structures such as latent spaces or latent class. I will focus on designing generative models able to reveal several clustering viewpoints and extend them to the deep-learning setting. I will also develop collaboration on these topics with doctors and retailers.

Clustering of various kind of data (recurrent event data, classifying subspaces, functional data categorical functional data)

How do you feel about joining the Institute?

I am very excited to join the Institute. I like the spirit of the Institute of both developing core elements of IA and the link between IA and the society such as health and smart territories. I think that this co-construction of IA with applications fields is essentials in order to adapt IA developed in laboratory with real life problems. Another exciting topic of the Institute is the dissemination of IA in all the society. I think that it will help in reducing the gap between the high expectation of the society and the real performances by introducing IA at the good level in human organisation.