news

Apr 07, 2026 Together with Andrew Wang, we will be presenting our library DeepInverse at the PyTorch Conference Europe 2026 in Paris on April 07, 2026. Please come and say hi if you are attending the conference!
Jan 04, 2026 New pre-print titled “An analytic theory of convolutional neural network inverse problems solvers”. We derive an analytical theory of convolutional neural network (CNN) inverse problems solvers: under 2 structural constraints, we found analytical, interpretable, and closed-form solutions to the CNN inverse problems solvers. We validate our theory on a large class of CNN architectures and inverse problems.
Nov 05, 2025 Our DeepInv library is now officially part of the Pytorch Ecosystem
Aug 04, 2025 New pre-print titled “How Diffusion Prior Landscapes Shape the Posterior in Blind Deconvolution”. We study the problem of blind deconvolution with state-of-the-art diffusion models as prior.
Mar 17, 2025 Presenting the poster of our work on Blind Deconvolution in Optical Microscopy at IABM 2025 and giving talk at IRIT Seminar.
Nov 27, 2024 Representing the DeepInv community to receive the Open Science Prize organized by the Ministère de l’enseignement supérieur et de la recherche.
Nov 24, 2024 New pre-print titled “Comparing Plug-and-Play and Unrolled networks”. We review their similarities and differences in a Bayesian framework and relates each method to a statistical estimator.
Sep 27, 2024 Participating to the DeepInv Hackathon at CIRM from 28-29 Nov. 2024.
Jun 16, 2024 Our paper titled “DeepVibes: Correcting Micro-Vibrations in Satellite Imaging with Pushbroom Cameras” has been accepted to be published in IEEE Transactions on Geoscience and Remote Sensing.