Corentin LEGER

ML Research Engineer at @instadeepai. Prev @Huawei, @FlowersInria

[Guillaume Pourcel @GuillaumeAP]

🧵Who said backprop was the only way to get gradients?
With Maxence Ernoult (@GoogleDeepMind), we investigated a hardware-friendly alternative that turns physical systems into self-learning machines: Recurrent Hamiltonian Echo Learning (RHEL)!
🔥RHEL @NeurIPS25 (oral)🔥

RHEL leverages the same physics for inference & learning. 🔁
Jump to the tweet [DEMO] for a simple interactive demo (no math 🙃, just intuitive mechanics ⚙️) of how RHEL turns a physical system into a self-learning machine 🕹️
📆 Meet us in person next week @EurIPSConf & @NeurIPSConf !!!
🇩🇰 Copenhagen Oral: Fri 5 Dec p.m CET
🇺🇲 San Diego
Oral: Fri 5 Dec 4:10 p.m. PST
Poster: Fri 5 Dec 4:30 p.m. PST
WiML Workshop Poster: Tue 2 Dec 6 pm PST (follow up neuro work for theorizing learning in the brain 🧠 presented by @AliceDauphin4)

[Abdelhakim Benechehab @abenechehab]

📢 New paper alert !!

How to use Policy Gradient methods without explicit rewards?

We address this question in our new work "From Data to Rewards: a Bilevel Optimization Perspective on Maximum Likelihood Estimation"

📜 https://t.co/mt0JhC7ZT7
🖥️ https://t.co/CtiqMyUt9q

1/🧵

[Clément Moulin-Frier @ClemMoulinFrier]

3 papers from @FlowersINRIA are presented this week at the @ALifeConf , leveraging curiosity-driven AI to explore complex behaviors in #Lenia

All first authors are attending the conference if you want to discuss with them

Link to interactive websites, papers and code below:

[Bartłomiej Cupiał @CupiaBart]

Almost all agentic pipelines prompt LLMs to explicitly plan before every action (ReAct), but turns out this isn't optimal for Multi-Step RL 🤔 Why?
In our new work we highlight a crucial issue with ReAct and show that we should make and follow plans instead🧵

[Ambroise Odonnat @ ICML 2026 @AmbroiseOdonnat]

🤔Why is tool use so effective for LLMs?

In our new work, we provide theoretical and empirical evidence that tool-augmented workflows are not just practical but also provably more scalable.

📜 https://t.co/iBP1EDnt8T
🖥️ https://t.co/rxOnExsop7

1/🧵

[Ambroise Odonnat @ ICML 2026 @AmbroiseOdonnat]

🚀 NeurIPS@Paris is back for a 5th edition at the SCAI. Meet us in central Paris to discuss recent advances in AI!
📆 25th & 26th Nov. 2025
🌐https://t.co/tgZJ1h2hIF
🎓 Committee: Chloé-Agathe Azencott, @BachFrancis, Claire Boyer, @gerardbiau , @VianneyPerchet , @jeanphi_vert

[Julien Pourcel @PourcelJulien]

Introducing SOAR 🚀, a self-improving framework for prog synth that alternates between search and learning (accepted to #ICML!)

It brings LLMs from just a few percent on ARC-AGI-1 up to 52%

We’re releasing the finetuned LLMs, a dataset of 5M generated programs and the code.

🧵

[Loris Gaven @LorisGaven]

🔔 Join our MAGELLAN talk on July 2!

We'll explore how LLM agents can monitor their own learning progress and choose what to learn next, like curious humans 🤔

1h presentation + 1h Q&A on autotelic agents & more!

📅 July 2, 4:30 PM CEST
🎟️ https://t.co/aYo8Xutm4q

[Clément ROMAC @ ICML 2025 @ClementRomac]

I'll be at RLDM this week to present our new paper: WorldLLM 😃

In the same spirit as our previous works — e.g. GLAM, MAGELLAN... — which investigate how to ground LLMs through interactions with external environments, WorldLLM takes a deep dive into the world modeling aspect.

[Loris Gaven @LorisGaven]

I'm happy to share that MAGELLAN has been accepted at #ICML2025! 🎉 See you in Vancouver!

[Pierre-Yves Oudeyer @pyoudeyer]

Generative AI is a cultural transmission technology:
it plays a growing role in generation, selection and transmission of ideas/opinions in human society 🧠🔄🌐

And yet we understand very little of this dynamics at this point 🤔❓

A step forward is our #ICLR2025 paper !👇

[Gaia Molinaro 🦋 gaiamolinaro.bsky.social @gaia_molinaro]

Check out this work today at 10 am @iclr_conf! Learn what's wrong with evaluating #LLMs after a single interaction, and how cultural attraction theory can help us do better. Poster #288 @Jeremy__Perez @pyoudeyer @MaximeDerex @KovacGrgur @cedcolas @corentinlger @ClemMoulinFrier

[Nicolas Yax @nicolas__yax]

🔥Our paper PhyloLM got accepted at ICLR 2025 !🔥
In this work we show how easy it can be to infer relationship between LLMs by constructing trees and to predict their performances and behavior at a very low cost with @StePalminteri and @pyoudeyer ! Here is a brief recap ⬇️

[Carta Thomas @CartaThomas2]

🚀 Introducing 🧭MAGELLAN—our new metacognitive framework for LLM agents! It predicts its own learning progress (LP) in vast natural language goal spaces, enabling efficient exploration of complex domains.🌍✨Learn more: 🔗 https://t.co/uGLBSsOgMn #OpenEndedLearning #LLM #RL

[Corentin LEGER @corentinlger]

@AdilZtn @FlowersINRIA thanks (:

[Corentin LEGER @corentinlger]

🚀 Want to simulate #AI/#ALife agents in a physics-driven world? Meet Vivarium🌱– a new multi-agent simulator from @FlowersINRIA , designed for research & education!
✅ Large-scale simulations with #JAX
✅ User-friendly web interface
✅ Real-time control with jupyter notebooks

[Corentin LEGER @corentinlger]

👏 Built with @Clement_MF_ and Martial Marzloff!

Explore Vivarium🌱here : https://t.co/BjMgmO0fyN

[Corentin LEGER @corentinlger]

🎓 Teaching multi-agent systems? Vivarium makes it easy!
✅ Ready-to-use Jupyter notebook sessions for hands-on learning
✅ Covers key concepts like reactive behaviors & environmental dynamics
✅ Used in CISC Master's @UPFBarcelona

📚 Explore them here: https://t.co/v5cobZZyoE

[Corentin LEGER @corentinlger]

🔮 What’s next for Vivarium?
- A more intuitive API is in development
- Soon, you’ll be able to modularly compose diverse dynamics – imagine Braitenberg Vehicles interacting with Particle Lenia in just a few lines of code!

🎥 Check out the demo: https://t.co/yAJSzMDIeg

[Corentin LEGER @corentinlger]

🎮 What can you do with Vivarium?
- Simulate agents with sensors & motors inspired by Braitenberg Vehicles
- Make them interact in a 2D physics world (built with JAX & JAX-MD)
- Visualize & control simulations dynamically in real time

🔗 Try it here: https://t.co/BjMgmO0fyN