Ph.D. student under the direction of R. COUILLET

Thesis title: Random matrix theory, a domain with various widesparead applications in mobile communications and lately in signal processing at large, is now in a state of maturity allowing one to handle more complex scenarios of deep importance to the BigData paradigm, especially those involving large dimensional graphs. In this thesis, we will study spectral clustering based methods for community detection in graphs as well as state changes in random graphs. The envisionned applications are numerous, so we will need to develop a whole framework of tools to handle these systematically. A second part of the thesis will deal with the study of echo-state neural networks, that are fundamentally based on the random nature of the underlying neural network. The outcome of the thesis is to allow to be a main driver of the random matrix research for large dimensional graphs and to be a leader in the related applications of these tools, thanks to the development of yet to be discovered original methods.


Journal papers

Conference papers