Michaël Defferrard

Twitter GitHub CV EPFL
photograph of me

Research

I am currently pursuing a PhD at the LTS2 Signal Processing laboratory of EPFL, under the supervision of Prof. Pierre Vandergheynst. My current research interest is to analyze and understand unstructured data, for example data lying on transportation, energy, social or brain networks. My goal is to generalize the success of Deep Learning beyond the traditional modalities such as images, sound, or videos, that are supported by n-dimensional Euclidean grids. I am currently developing Convolutional Neural Networks on data structured by arbitrary graphs, which are versatile representations of heterogeneous pairwise relationships, by leveraging the mathematical tools developed in spectral graph theory and the emerging field of signal processing on graphs.

Besides research, I am teaching practical Data Science. Either by preparing and giving exercises for a master course at EPFL or a block course targeted at industrial practitioner. I used to maintain a digital lab notebook with some notes as well as relevant tools, datasets, groups, events or courses.

Before starting a PhD in September 2015 I obtained a MSc in Electrical and Electronic Engineering at EPFL with a focus in Information Technologies and a minor in Computational Neurosciences. During that time I was a research assistant at LTS2 as well as a software engineer at Infoteam. Before that I obtained my BSc in an engineering school after a vocational training as an electronics specialist. See my academic CV or LinkedIn for more details on my education or professional experience.

Publications

Google Scholar arXiv EPFL Infoscience

Selected projects

Private

Besides scientific research, I play brass band music at the Brass Band Fribourg and my town's band. I enjoy to read, some of which is on Goodreads, and to travel. I post about some of my life on Facebook. I sometimes answer questions on Stack Overflow and Quora, and participate to Hacker News and reddit.