Irène Waldspurger

I have been a PhD student in the DATA team (École normale supérieure, computer science department), under the supervision of Stéphane Mallat. I was at the same time a teaching assistant in the department of mathematics.

I am now a postdoctoral fellow at MIT Institute for Data, Systems, and Society, working with Philippe Rigollet in the Center for Statistics.


During my PhD, I mainly worked on inverse problems, especially on phase retrieval. Using harmonic analysis tools, I studied the phase retrieval problem for the wavelet transform from a theoretical point of view. I also worked on two numerical reconstruction algorithms. The first one uses convexification techniques and can be applied to any phase retrieval problem. The second one applies only to the case of the wavelet transform, but is much faster.

I am also interested in machine learning, classification and detection problems. I studied the mathematical properties of the scattering transform and, in a preliminary manner, its link with the convolutional networks used in computer vision, and the role of invariance to small diffeomorphisms for image analysis or classification.

Curriculum vitae

PhD thesis


Teaching / Enseignement