Skip to content. | Skip to navigation

Personal tools

Sections

UMR 5672

logo de l'ENS de Lyon
logo du CNRS
You are here: Home / Seminars / Machine Learning and Signal Processing / Random matrices in optics: optical computing and computational imaging

Random matrices in optics: optical computing and computational imaging

Jonathan Dong (post-doc EPFL)
When Mar 19, 2024
from 01:00 to 02:00
Attendees Jonathan Dong
Add event to calendar vCal
iCal

 

Jonathan Dong

Title: Random matrices in optics: optical computing and computational imaging

Abstract: Random matrices are ubiquitous in both optics and statistical physics. They serve as powerful tools to model wave propagation in complex media and are associatd to the universal phenomena and phase transitions in statistical physics. In this presentation, we explore their applications in two nonlinear problems: optical computing and computational imaging. Optical computing harnesses light to execute light to perform computation. We will show how to implement feedforward and recurrent neural networks in an optical system, and study the mean-field limit of the associated random networks. In computational imaging, we focus on the phase retrieval problem, a quadratic equation to solve in order to retrieve images. We will show how tools from statistical physics shed insights on solvability, which algorithm to use and broader applicability beyond the random setting. 

Website: https://jonathandong.fr/

In Room M7.101 of Monod campus, ENSL.