Tessera
Folgen

Julian Tachella

@tachellajulian.bsky.social

CNRS research scientist, based at ENS de Lyon I'm interested in AI for imaging inverse problems Looking to hire phds/postdocs! 🇦🇷🇬🇧🇫🇷 Website: https://tachella.github.io/ Deepinverter: https://deepinv.github.io/

79 Posts279 Follower136 Folgt
Julian Tachella@tachellajulian.bsky.social · 9mo ago

📖 We put together with Mike Davies a review of self-supervised learning for inverse problems, covering the main approaches in the literature with a unified notation and analysis. arxiv.org/abs/2601.03244

Self-Supervised Learning from Noisy and Incomplete Data

Many important problems in science and engineering involve inferring a signal from noisy and/or incomplete observations, where the observation process is known. Historically, this problem has been tac...

arxiv.org

Julian Tachella@tachellajulian.bsky.social · last yr.

☀️ Just wrapped up the DeepInverse Hackathon! We had 30+ imaging scientists from all over the world coding during 3 days next to the beautiful Calanques in Marseille, France. It was a great moment to meet new people, discuss science, and code new imaging algorithms!

Bild
Julian Tachella@tachellajulian.bsky.social · last yr.

🚀🚀 Reconstruct-Anything-Model: A single model to rule all imaging tasks! We challenge current beliefs and show that a single U-Net can obtain impressive performance across a wide variety of tasks, **without** relying on expensive iterative schemes such as unrolling, PnP, diffusion

Julian Tachella@tachellajulian.bsky.social · last yr.

Deriving Tweedie's formula from SURE let's you also derive other interesting generalisations, such as a formula that doesn't require knowledge about the noise level, and it can be generalized well beyond isotropic gaussian noise.

Peyman Milanfar@docmilanfar.bsky.social · 2y ago

One more thing: Most people I've seen recently derive Tweedie from Stein's Unbiased Risk Estimator (SURE). I don't get that - it's messy and requires much stronger regularity conditions. I mean - Tweedie invented the formula before Stein invented SURE after all! @gabrielpeyre.bsky.social FYI

Julian Tachella@tachellajulian.bsky.social · 2y ago

🚀 Do you want to learn about self-supervised learning for inverse problems? ▶️ Check out the 3-hour tutorial presented by Mike Davies (University of Edinburgh) and myself (CNRS, ENS Lyon), covering all the recent advances in a unified and simplified framework! youtube.com/playlist?lis...

Self-Supervised Learning for Inverse Problems - YouTube

Tutorial by Julián Tachella (CNRS, ENS Lyon) & Mike Davies (University of Edinburgh) given at the University of Edinburgh, February 2025.I. "Introduction"II....

youtube.com

Julian Tachella@tachellajulian.bsky.social · 2y ago

🔉New paper "Generalized Recorrupted-to-Recorrupted" 🔉 with @bemc22.bsky.social and Jorge Bacca - Generalizes R2R self-supervised loss for noise belonging to the exponential family. - Shows asymptotic equivalence with SURE. paper: arxiv.org/abs/2412.04648 code: github.com/bemc22/Gener...

Generalized Recorrupted-to-Recorrupted: Self-Supervised Learning Beyond Gaussian Noise

Recorrupted-to-Recorrupted (R2R) has emerged as a methodology for training deep networks for image restoration in a self-supervised manner from noisy measurement data alone, demonstrating equivalence ...

arxiv.org

Jon Tamir@jtsense.bsky.social · 2y ago

Hi -- I have a postdoc position open in Computational MRI -- looking for experience and interest in MRI acquisition (in addition to reconstruction). Offer competitive pay. Opportunity to collaborate with Center for GenAI and Dell Med School. Bonus if you get your app in before Dec 1!