Link prediction in random graphs

Quentin Duchemin

Labo. d'Analyse et de Mathématiques Appliquées, Université Gustave Eiffel

Nowadays, random graph models are widely used to extract relevant information on complex systems. One task of particular interest is to infer connections in networks where we only observe a partial amount of all possible links. In this talk, I will present one random graph model that allows to tackle such problems in a high dimensional setting. The talk is designed to be accessible to every PhD students coming from any field of maths. The purpose is mainly to introduce some key concepts/questions/tools of high dimensional statistics and optimization.