Gilles Louppe | Lightnews

Light up
your news

About Privacy Terms Help

Gilles Louppe

Gilles Louppe

@glouppe.bsky.social

3.8K followers 930 following 78 posts

AI for Science, deep generative models, inverse problems. Professor of AI and deep learning @universitedeliege.bsky.social. Previously @CERN, @nyuniversity. https://glouppe.github.io

Computer science 62%

Physics 22%

more

Posts Replies Media Videos

Pinned

Gilles Louppe @glouppe.bsky.social · Apr 29

<proud advisor>
Hot off the arXiv! 🦬 "Appa: Bending Weather Dynamics with Latent Diffusion Models for Global Data Assimilation" 🌍 Appa is our novel 1.5B-parameter probabilistic weather model that unifies reanalysis, filtering, and forecasting in a single framework. A thread 🧵

Reposted by Joanna Bryson, Gilles Louppe

sakanaai.bsky.social @sakanaai.bsky.social · 5h

Introducing DroPE: Extending Context by Dropping Positional Embeddings

We found embeddings like RoPE aid training but bottleneck long-sequence generalization. Our solution’s simple: treat them as a temporary training scaffold, not a permanent necessity.

arxiv.org/abs/2512.12167
pub.sakana.ai/DroPE

Reposted by Alex Mesoudi, Gilles Louppe

Mason Youngblood @masonyoungblood.bsky.social · 3d

Proc B with @sampassmore.bsky.social! We used simulations to explore the innovation strategies of speed climbers 🧗‍♀️ Innovation is higher among slower athletes and lower when the population size is larger, and the overall balance of innovation and copying appears to be suboptimal 🔗 bit.ly/499QjZM

Simulation-based inference with deep learning suggests speed climbers combine innovation and copying to improve performance

Abstract. In the Olympic sport of speed climbing, athletes compete to reach the top of a 15 m wall as quickly as possible. Since the standardization of the

Reposted by Gilles Louppe

Andrew Gordon Wilson @andrewgwils.bsky.social · 4d

We introduce epiplexity, a new measure of information that provides a foundation for how to select, generate, or transform data for learning systems. We have been working on this for almost 2 years, and I cannot contain my excitement! arxiv.org/abs/2601.03220 1/7

Reposted by Gilles Louppe

AI Firehose @ai-firehose.column.social · 5d

A study shows that using AI tools like LLMs lowers cognitive engagement and critical thinking in essay writing, signaling a trade-off between convenience and growth. LLM users had weaker memory recall and ownership, raising vital talks about AI's role in education. https://arxiv.org/abs/2506.08872

Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task

ArXiv link for Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task

Reposted by Gilles Louppe

Shubhendu Trivedi @shubhendu.bsky.social · 8d

I feel like it is far more important to teach SVMs, and the machinery around them, in basic ML courses now than it was in 2012.

Reposted by Paul Matthews, Michael Müller, Iñaki Aldasoro , and 2 more Paul Matthews, Michael Müller, Iñaki Aldasoro, Greg Linden, Gilles Louppe

Melanie Mitchell @melaniemitchell.bsky.social · 10d

2020s: AI "System Prompts" are lengthy, carefully constructed sets of expert rules about a particular domain, created by "prompt engineers".

1980s: AI "Expert Systems" were lengthy, carefully constructed sets of expert rules about a particular domain, created by "knowledge engineers".

Reposted by Gilles Louppe

Johann Brehmer @johannbrehmer.bsky.social · 10d

Happy New Year!

Do your plans for 2026 include...
- working with a great team lead by @wellingmax.bsky.social and @aronwalsh.github.io?
- living in Amsterdam, Berlin, London, or Cambridge?
- using fun tools from ML and material science?
- solving important problems?

Then join us at CuspAI!

1/2

Gilles Louppe @glouppe.bsky.social · 12d

I just started playing with Claude Code too. It is really impressive 🤯

Reposted by Gilles Louppe

Science X / Phys.org @sciencex.bsky.social · 13d

A biologically detailed brain model has matched animal learning performance and uncovered previously overlooked neuron activity linked to errors during visual categorization tasks. doi.org/hbhcsn

Biology-inspired brain model matches animal learning and reveals overlooked neuron activity

A new computational model of the brain based closely on its biology and physiology has not only learned a simple visual category learning task exactly as well as lab animals, but even enabled the discovery of counterintuitive activity by a group of neurons that researchers working with animals to perform the same task had not noticed in their data before, reports a team of scientists at Dartmouth College, MIT, and the State University of New York at Stony Brook.

Gilles Louppe @glouppe.bsky.social · 14d

I have been in AI for science for 10+ years now. Real benefits, real impact. But 'AI' feels like a bad word now, with the media running 10:1 negative. Yet almost everyone around me uses LLMs daily. Honestly lost.

Reposted by Gilles Louppe

Ethan Mollick @emollick.bsky.social · 15d

Coding.

Reposted by Gilles Louppe

Nicola Branchini @nicolabranchini.bsky.social · 15d

Very nice arxiv.org/abs/2512.20003

Control Variate Score Matching for Diffusion Models

Diffusion models offer a robust framework for sampling from unnormalized probability densities, which requires accurately estimating the score of the noise-perturbed target distribution. While the sta...

Reposted by Gilles Louppe

Luigi Acerbi @lacerbi.bsky.social · 16d

What an amazing Yule gift from @stefanradev.bsky.social & colleagues: a tour-de-force tutorial on diffusion models for simulator-based inference.

This is one of the most comprehensive and useful review/tutorials I have ever seen -- a must read! Kudos to all the authors!

arxiv.org/abs/2512.20685

Diffusion Models in Simulation-Based Inference: A Tutorial Review

Diffusion models have recently emerged as powerful learners for simulation-based inference (SBI), enabling fast and accurate estimation of latent parameters from simulated and real data. Their score-b...

Reposted by Gilles Louppe

Flatiron Institute @flatironinstitute.org · 23d

New AI models from @polymathicai.bsky.social use knowledge from scientific fields and apply it to others. Walrus (seen below) can tackle systems from exploding stars to Wi-Fi signals: https://www.simonsfoundation.org/these-new-ai-models-are-trained-on-physics-not-words-and-theyre-driving-discovery/

Reposted by Gilles Louppe, Michael Larkin

Itai Yanai @itaiyanai.bsky.social · 23d

Doing a PhD is - at heart - one long discussion with your mentor. The discussion changes over time - with unexpected turns and ups & downs - but through it all is a pair of people discussing a topic endlessly to make sense of it.
PhD students: choose someone you like to talk to!

Reposted by Katharine Hayhoe, Gilles Louppe

Aida Alvera-Azcárate @aida-alvera.bsky.social · 25d

Your pretty picture of the day 😀 🌊with of course many interesting processes! Water rich in suspended sediments is moved around by ocean currents in the Bay of Biscay. I'm curious about the long-range waves seen in the dark region, can't decide if these are oceanic, atmospheric, or an artifact 😅

Satellite image taken by Sentinel 2 over the Bay of Biscay on 15 December 2025, showing waters full of suspended sediments in the top right half of the image, in turquoise colour, and open waters in dark blue in the low left half of the image. Image taken from https://browser.dataspace.copernicus.eu

Reposted by Gilles Louppe

CIERA @nuciera.bsky.social · Dec 11

A discovery fit for Star Wars: CIERA astronomers have directly imaged a Tatooine-like planet orbiting two suns — and it’s closer to its stars than any other directly imaged planet in a binary system. 🌟🌟🪐https://bit.ly/4q8Dfto
@thejason.wang #Exoplanets #Astronomy #Northwestern #CIERA

Reposted by Gilles Louppe

Percy Balemans 📚 @creative-xl8.bsky.social · Dec 10

"Machine translation would handle the heavy lifting, and we’d add the expert polish. Except the machine made a mess. And we’re doing expert-level cognitive labour to clean it up for cleaning wages.

We’re not mops. We’re translators."

#translation #xl8

www.linkedin.com/pulse/open-l...

An open letter to colleagues: we are not mops

After more than two decades working in legal translation across four languages and multiple jurisdictions, I’ve watched our profession reshaped by forces that claim efficiency as a path to our expenda...

www.linkedin.com

Reposted by Gilles Louppe

Kosta Derpanis @csprofkgd.bsky.social · Dec 8

Repeat after me: This is NOT a paper contribution, it’s an expected component!

Reposted by Carl T. Bergstrom, Veli‐Matti Karhulahti, Gilles Louppe

Julian Togelius @togelius.bsky.social · Dec 8

was at an event on AI for science yesterday, a panel discussion here at NeurIPS. The panelists discussed how they plan to replace humans at all levels in the scientific process. So I stood up and protested that what they are doing is evil.

Full post:
togelius.blogspot.com/2025/12/plea...

Please, don't automate science!

I was at an event on AI for science yesterday, a panel discussion here at NeurIPS. The panelists discussed how they plan to replace humans a...

togelius.blogspot.com

Reposted by Gilles Louppe

julienbrandoit.bsky.social @julienbrandoit.bsky.social · Dec 7

Why can neurons with completely different ion channels fire identically? 🧠

Neuronal degeneracy isn't a bug - it's a feature. Come see how we reconstruct entire degenerate populations from spike times within an interpretable feedback framework.

Poster @dataonbrainmind.bsky.social

#NeurIPS2025 #DBM

Reposted by Gilles Louppe

Gérôme Andry @Neurips2025 @gandry.bsky.social · Dec 6

If you are interested in diffusion models for dynamical systems and posterior inference, stop by our poster at ML4PS workshop (@neuripsconf.bsky.social) at 11am!

Come and see how we apply modern methods to the challenging field of global weather inference!

Reposted by Stephen M. Walt, Anna O. Law, Peter M. Cox , and 29 more Stephen M. Walt, Anna O. Law, Peter M. Cox, Will Jennings, Peter Thorne, Luc Int Panis, Albert Weale, Tim Bale, Rosemary A. Joyce, Conor McCarthy, Rupert Read, Alison Phipps, Edgar Morgenroth, Göran Bolin, Smith, Beatriz Gallardo Paúls, Rasmus K. Storm, Laurence Piper, Jason Lyall, Johannes Breuer, Jos Akkermans, Martin Lenz, Tim Stephens, Gilles Louppe, Olivier Mannoni, Nicolas Hervé, Manisha Sinha, Nathan P. Kalmoe, Stefano Gualeni, Morwenna Ludlow, Antonio Fatás, David Darmofal

derek guy @dieworkwear.bsky.social · Dec 6

The Nobel Prize committee should announce the World Cup winner tomorrow

Reposted by Gilles Louppe

Kyle Cranmer @kylecranmer.bsky.social · Dec 5

Later we did early work using message passing graph NNs with @joanbruna.bsky.social and @johannbrehmer.bsky.social. Fully connected message passing NNs are pretty similar to transformers, but this was in 2017.
ml4physicalsciences.github.io/2017/files/n...

ml4physicalsciences.github.io

Reposted by Gilles Louppe

Kyle Cranmer @kylecranmer.bsky.social · Dec 5

An unexpected shoutout from @kyunghyuncho.bsky.social in his NeurIPS keynote. @glouppe.bsky.social & I collaborated with him to bring ideas from natural language to high-energy physics back in 2016.
arxiv.org/abs/1702.00748

Reposted by Gilles Louppe

Martin Trapp @trappmartin.bsky.social · Dec 5

Beautiful keynote by @gaelvaroquaux.bsky.social at @euripsconf.bsky.social .

Reposted by Gilles Louppe

Gaël Varoquaux @gaelvaroquaux.bsky.social · Dec 4

Super excited (and a bit nervous) to give a keynote tomorrow afternoon at #EurIPS. I'll be math, LLMs, and politics 🔥

Reposted by Gilles Louppe

Annalena Kofler @annalenakofler.bsky.social · Dec 3

1/ 🌀 New paper alert! We introduce Dingo-T1, a flexible transformer-based deep learning model for gravitational-wave (GW) data analysis. It adapts to different detector & frequency settings, improving inference efficiency and flexibility

🚀 #AI #MachineLearning #Physics #Astronomy #AcademicSky

Scientific poster with dark background and two black holes illustrated in the center. The paper visualizes gravitational waves, and explains how parameter estimation is performed with DINGO. The standard DINGO model and the DINGO-T1 architectures are illustrated and results are shown. For example, it is possible to reanalyze the same event with different detector configurations with DINGO-T1, illustrated bz a corner plot.

Reposted by Gilles Louppe

Sung Kim @sungkim.bsky.social · Dec 3

What coding with an LLM feels like sometimes.

Reposted by Gilles Louppe

Chaitanya K. Joshi @chaitjo.bsky.social · Dec 3

Introducing gRNAde: our own little "AlphaGo Moment" for RNA design! 🧬🚀

📝: tinyurl.com/gRNAde-paper

Unlike proteins, RNA design has long relied on "wisdom of the crowd" (human experts) or the slow crawl of directed evolution — gRNAde changes that! 🧵👇