Tessera
Folgen

Emmanuel Soubies

@esoubies.bsky.social

CNRS researcher at IRIT. Computational Imaging, Microscopy, Optimization, Machine Learning Website: https://www.irit.fr/~Emmanuel.Soubies/

4 Posts237 Follower195 Folgt
Jonny@jchirinosrguez.bsky.social · last mo.

New paper out! We study a novel property ensuring linear convergente of Bregman Proximal Gradient Methods, and show the applicability of our results to KL regression problems, present for instance in Poisson inverse problems. 🖤 (With @esoubies.bsky.social, Christian Daniele, and Cédric Févotte)

arxiv math.OC@arxiv-math-oc.bsky.social · last mo.

Jonathan Chirinos-Rodr\'iguez, Christian Daniele, C\'edric F\'evotte, Emmanuel Soubies On The Linear Convergence of Bregman Proximal Gradient Methods with Applications to Kullback--Leibler regression https://arxiv.org/abs/2607.05539

Samuel Vaiter@samuelvaiter.com · 8mo ago

Les inscriptions aux journées MODE 2026 à Nice sont désormais ouvertes. Elles se dérouleront du 18 au 20 mars à l'Hôtel Saint-Paul. Les inscriptions sont ouvertes jusqu'au 1 mars (majoration > 9/02). La deadline pour soumettre une communication est le **15 janvier**.

Bild
Jonny@jchirinosrguez.bsky.social · 9mo ago

New paper out! Together with my postdoc advisors @esoubies.bsky.social and Cédric Févotte :) We study the optimization landscape of L0-Bregman relaxations, which naturally arise when one optimizes non-quadratic fidelity terms coupled with the L0-norm. 1/n

arxiv math.OC@arxiv-math-oc.bsky.social · 9mo ago

Jonathan Chirinos-Rodriguez, C\'edric F\'evotte, Emmanuel Soubies Optimization landscape of $\ell_0$-Bregman relaxations https://arxiv.org/abs/2511.12157

Mathurin Massias@mathurinmassias.bsky.social · last yr.

Then why does flow matching generalize?? Because it fails! The inductive bias of the neural network prevents from perfectly learning u* and overfitting. In particular neural networks fail to learn the velocity field for two particular time values. See the paper for a finer analysis 😀

Bild
Mathurin Massias@mathurinmassias.bsky.social · last yr.

It was received quite enthusiastically here so time to share it again!!! Our #ICLR2025 blog post on Flow Matching was published yesterday : iclr-blogposts.github.io/2025/blog/co... My PhD student @annegnx.bsky.social will present it tomorrow in ICLR, 👉poster session 4, 3 pm, #549 in Hall 3/2B 👈

A Visual Dive into Conditional Flow Matching | ICLR Blogposts 2025

Conditional flow matching (CFM) was introduced by three simultaneous papers at ICLR 2023, through different approaches (conditional matching, rectifying flows and stochastic interpolants). <br/> The m...

iclr-blogposts.github.io

Christophe 🔬 Leterrier@christlet.bsky.social · last yr.

We have a new review out! In this Microscopy & Microanalysis piece, Louisa Mezache and I survey some of the latest commercially available microscope technologies: Nikon NSPARC, Re-scan Gaia, CSR Biotech MI-SIM, refinements to STED, SMLM from Bruker and Abbelight… Check it out with free access:

Advancing Super-Resolution Microscopy: Recent Innovations in Commercial Instruments

Abstract. Super-resolution microscopy techniques have accelerated scientific progress, enabling researchers to explore cellular structures and dynamics wit

academic.oup.com

Sandrine Lévêque Fort 🔬@slfort.bsky.social · 2y ago

If you are interested by building or applying new microscopes around single molecule, FLIM, in depth imaging, live-cell+ SMLM, we are opening post-doc positions for various backgrounds funded by ERC grant, don’t hesitate to contact me (flexible start). shorturl.at/fYNij shorturl.at/gDQkz

Portail Emploi CNRS - Offre d'emploi - Post-doctorant (H/F) pour le développement de nouveaux concepts en microscopie de fluorescence super-résolue pour l'observation de cellules vivantes

shorturl.at

Paul Hagemann @yungbayesian.bsky.social · 2y ago

Our paper "PnP-Flow: Plug-and-Play Image Restoration with Flow Matching" has been accepted to ICLR 2025. Here a short explainer: We want to restore images (i.e., solve inverse problems) using pretrained velocity fields from flow matching. However, using change of variables is super costly.

Bild
Laurent Duval (research amateur)@laurentduval.bsky.social · 2y ago

Have some curiosity on investigating the well-known PCA (Principal Component Analysis) with a glimpse of sparse optimization? With homogeneous or scale-invariant SOOT/SPOQ norm ratio penalties? Join INRIA for a Master Internship in 2025 (INRIA) Revisiting PCA with norm-ratio sparsity penalties

Bild