Very proud of my colleague @etnlalib, his group, collaborators and my own students who, as a part of team Limelight have won a ($10M) XPrize - focused on the goal of increasing our understanding of the Amazon Rainforest #XPRIZERainforest & thanks to @IVADO_Qc for the support !

In-flight weight updates have gone from a “weird trick” to a must to train LLMs with RL in the last few weeks. If you want to understand the on-policy and throughput benefits here’s the CoLM talk @DBahdanau and I gave:

Generative Point Tracking with Flow Matching My latest project with @AdamWHarley @CSProfKGD @DerekRenderling @chrisjpal Project page: https://t.co/cs4zFEuLYU Paper: https://t.co/sa9NdFlOgP Code: https://t.co/F4Ug3JWkRX

✨ What if we could tune Frontier LLM agents without touching any weights? Meet JEF-Hinter, an agent capable of analyzing multiple offline trajectories to extract auditable and timely hints💡 In our new paper 📄 , we show significant performance gains on downstream tasks ⚡ with

We did lots of good work since PipelineRL release in May: ⚙️ higher throughput, seq parallel training, multimodal, agentic RL 📜 white paper with great explanations and results: https://t.co/F3YsIbNRUy We'll present today at CoLM EXPO, room 524C, 1pm!

Very excited to be presenting Pipeline RL this afternoon at CoLM. Join us if you are interested in fast on policy RL training for LLMs 🚀

New paper 📜: Tiny Recursion Model (TRM) is a recursive reasoning approach with a tiny 7M parameters neural network that obtains 45% on ARC-AGI-1 and 8% on ARC-AGI-2, beating most LLMs. Blog: https://t.co/w5ZDsHDDPE Code: https://t.co/7UgKuD9Yll Paper:

With recursive reasoning, it turns out that “less is more”. A tiny model pretrained from scratch, recursing on itself and updating its answers over time, can achieve a lot without breaking the bank.

Follow Mila researchers on day 2 of #COLM2025. Full schedule of Mila-affiliated presentations here https://t.co/e4298p390Y

Suivez les chercheur·euse·s de Mila au jour 2 de la #COLM2025. Programme complet disponible ici https://t.co/JXvo6ASlhG

with yours truly

Huan Sun (@hhsun1) on improving safety of AI agents -- capabilities and safety should go hand-in-hand, not an afterthought -- definition: safety is unintentional alignment, and security involves adversarial attacks. For this talk, safety also involves security. -- Safety has to

SLAM Labs presents Apriel-1.5-15B-Thinker 🚀 An open-weights multimodal reasoning model that hits frontier-level performance with just a fraction of the compute.

@alexpiche_ Perfect time for for @OpenAI to talk about reducing hallucinations! Cause you know, our work on learning to abstain by iterative self-reflection with @alexpiche_ @amilios and @chrisjpal has just got accepted at TMLR 😉

Glad to see OpenAI prioritizing abstention responses in their paper! That's a great intro to our TMLR paper in which we developed an iterative self-reflection method for LLM to know when to abstain without ground truth and no additional cost at test time. https://t.co/xwNT68ejqm

UI-Vision vs GPT-5: Still holding the crown 👑 and far from saturation. GPT-5 has strengths in coding and reasoning, but when it comes to computer-use tasks, it’s still awkward to rely on it alone. And our team's UI-Vision (ICML 2025) remains a key and still unbeaten multimodal

We built a new 𝗮𝘂𝘁𝗼𝗿𝗲𝗴𝗿𝗲𝘀𝘀𝗶𝘃𝗲 + 𝗥𝗟 image editing model using a strong verifier — and it beats SOTA diffusion baselines using 5× less data. 🔥 𝗘𝗔𝗥𝗟: a simple, scalable RL pipeline for high-quality, controllable edits. 🧵1/

I think listening to this, with an amplifier set to 11, would be appropriate today.

Finally finished a blog post I've been working on (on and off) for months. It builds on a TMLR paper I published last year in model-based optimisation, but I wanted to explain things more clearly this time. More honest, more readable, more reflective. https://t.co/BBlaBDNUAP

We’re releasing SelvaBox, the largest tropical tree detection dataset from drone imagery. Our models trained on SelvaBox achieve competitive zero-shot detection performance on unseen tropical tree crown datasets, matching or exceeding competing methods. https://t.co/GGNoRdqHtw

In December 2024 @pankaj @gilad @willhorn and I put out a rather cryptic arXiv paper musing about the future of search: https://t.co/CfpU5E2HxN. I’m now able to share what I’ve been up to! 🧵(1/9)