Communication - Event

Workshop LARGR Lille 2023: conferences on statistical learning for graphs

09 to 10 Mar. 2023
Location :
Centre Inria de l'Université de Lille

40 avenue Halley , 59650 Villeneuve d'Ascq

Changed on 20/02/2023
Hemant Tyagi and Christophe Biernacki (MODAL project-team) are organising a two-day workshop on statistical learning for LARge scale GRaphs (LARGR), on March 9 and 10, 2023, in the amphitheatre of building B of the Inria centre of the University of Lille.

This workshop will cover topics relevant to statistical learning problems on large-scale networks, in particular, related to clustering, ranking, graph matching, graph neural networks (GNN), graph signal processing, and dynamic processes on networks.

It will include a combination of 6 long talks (1 hour each, by leading invited speakers in their fields, listed below) and 15 shorter talks (20 minutes each, mainly given by Ph.D. students/post-docs/young researchers). All talks will include a general introduction to make them accessible to a wide audience, not necessarily specialists in the field.

Guest speakers:

  • Marc Barthélemy (Université Paris-Saclay, CEA, CNRS)
  • Pierre Borgnat (CNRS, ENS Lyon)
  • Julien Hendrickx (UCLouvain)
  • Marc Lelarge (Inria et ENS Paris)
  • Pietro Liò (Université de Cambridge)
  • Catherine Matias (CNRS, Sorbonne Université)

Why should I be interested in the subject of "learning about graphs"?

In many problems in science and engineering, we access data in the form of pairwise relationships between a set of n objects. These pairwise relationships naturally lead to an underlying graph, with nodes corresponding to the objects, and edges encoding the pairs of objects for which information is available. The objective is then to learn an underlying latent structure associated with the objects using the available pairwise data.

Such problems arise in a wide range of applications, such as computer vision, recommender systems, sports tournaments, biology, and social sciences, to name a few. In many cases, the data is large with n in the thousands or even millions (e.g. social networks), making the algorithmic task practically difficult. The last decade has seen an impressive range of results - both from a theoretical and practical point of view - with several fundamental results, as well as recent breakthroughs made by Graph Neural Networks in many important applications.



Aim of the workshop

This workshop will bring together researchers from different communities in mathematics, statistics, computer science, and signal processing, and will provide a stimulating environment for interdisciplinary discussions. It will be of interest to both theorists and practitioners.

The objective of this workshop is threefold. The main objective is to gain a better understanding of some of the recent advances in this field - both in theory and in practice. Secondly, we seek to identify interesting and relevant avenues for future research. Finally, our aim is to disseminate information among the participants, and we hope to cultivate interest in the broad field of 'network science' among a wider audience.