• Country:

  Belgium

• Education:

  University of Mons

Valentin Leplat received engineering degrees in mechanical engineering fr om the Gramme Institute in Liège, Belgium, in 2012, and in computer science and applied mathematics engineering from the University of Mons, Belgium, in 2017. He worked for six years as an aerospace engineer at SONACA in Gosselies, Belgium. His main activities involved the numerical analysis (aerodynamics and structural) of aircraft and spacecraft structures. He participated in the design and certification of the wings of the CS-100 Bombardier, EMBRAER 190E2, and EMBRAER 175E2 aircraft. He completed his Ph.D. in applied mathematics in January 2021 on the topic of Nonnegative Matrix Factorizations (NMF), associated with the Department of Mathematics and Operations Research at the University of Mons, Belgium, under the supervision of Nicolas Gillis and Xavier Siebert. His main results are threefold: (1) the analysis of the geometry behind NMF, which allowed for the proposal of new models and efficient algorithms capable of recovering and identifying the factors that gave rise to the data (also referred to as the problem of uniqueness of the factorization); (2) the algorithmic aspect of NMF; in fact, computing a nonnegative factorization requires the resolution of challenging non-linear non-convex optimization problems under constraints, for which new optimization frameworks have been proposed; and (3) the first conic formulation of NMF, which allows for the simultaneous update of all factors of the decomposition instead of using block-coordinate schemes. In 2021, he was a postdoctoral research associate in applied mathematics at Université Catholique de Louvain (Belgium) under the supervision of Yurii Nesterov, wh ere he studied the complexity of some variants of NMF and derived optimization methods for minimizing concave functions over convex sets. Between October 2021 and July 2024, he worked at Skoltech (Moscow) as a senior research scientist and later as a leading research scientist under the supervision of Ivan Oseledets and Anh Huy Phan on the topics of machine learning, stochastic optimization, and tensor decompositions. Since July 2024, he joined Innopolis University as Assistant Professor among the Faculty of Computer Science and Engineering and the Institute of Data Analysis and Artificial Intelligence. His current research activities revolve around the development of efficient optimization methods for solving large-scale non-convex problems in both deterministic and stochastic settings, the development of randomized methods to compute tensor decompositions, and the study of modern tensor decomposition models (such as deep Nonnegative Matrix Factorization) with an emphasis on the problems of automatic model order selection, robustness, and identifiability.