Ponente
Descripción
ue to the significant increase of communications between individuals via social media (Facebook, Twitter, Linkedin) or electronic formats (email, web, e-publication) in the past two decades, network analysis has become an unavoidable discipline. Many random graph models have been proposed to extract information from networks based on person-to-person links only, without taking into account information on the contents. This talk will describe first the stochastic topic block model (STBM), a probabilistic model for networks with textual edges. We will address the problem of discovering meaningful clusters of vertices that are coherent from both the network topology and the text contents. Then, a classification variational expectation-maximization (C-VEM) algorithm will be described to perform inference. This work is supported by CNRS, INRIA, PIA, and the CARE-COVID-19 committee. It led to the development of the Linkage.fr platform which will be presented. The talk will describe new developments that we are working on with the use of variational auto-encoders and deep probabilistic graphical models. Finally, I will describe how we used Linkage to analyse the last French presidential election with a team of researchers and journalists from LeMonde.
