4/9 NeurIPS submissions accepted! Congrats to James Allingham, Bruno Mlodozeniec, Jose Miguel Hernández-Lobato, Luke Marks, Fazl Barez, Ryan Greenblatt, Fabien Roger, Dima Krasheninnikov, Stephen Chung, Scott Niekum and all of my other co-authors on these works! Summaries / links in thread...

Want to join Mila - Institut québécois d'IA an inspiring and vibrant ML research community? Submit your supervision request by December 1!

Podem encontrar o meu comentário mais detalhado aqui: publico.pt/2024/10/08/cie…

Multi-Agent RL fails in real life. Agents cooperating to solve tasks remains a utopia. -No scalable algorithms for general-sum games. -In a simple apple-harvesting game, PPO agents overharvest and ruin bushes. Advantage Alignment (ICLR 2025 Oral📢) is a huge step forward. 1/n

Tomorrow Sunday, 14h45! At ICLR 2026 2025 Workshop on Human-AI Coevolution (HAIC) Mila - Institut québécois d'IA

Thank you to the HAIC organizers ICLR 2026 ! Very friendly environment and interesting discussions today Mila - Institut québécois d'IA

🚨Today is the last day to apply to the Lisbon Machine Learning School! Hear from top speakers in the field and get to meet new people during a fantastic week in Lisbon. ‼️Apply today: tinyurl.com/mre5as8f

Interested in NNs with improved tail performance? Our #AISTATS2025 paper introduces Feasible Learning, a sample-centric learning paradigm that bounds per-sample losses! Nacho Hounie Juan Elenter Jose Gallego-Posada Meraj A. Ribeiro Simon Lacoste-Julien arxiv.org/abs/2501.14912 🧵

Check our poster tomorrow 3-6pm! AISTATS Conference

🥳 Just managed to finish the first draft of this work after almost four years of work 😅 (Im)possibility of Collective Intelligence arxiv.org/abs/2206.02786 Any feedback is highly appreciated 🙏

Meet Cooper — a PyTorch library for constrained optimization in deep learning. 🚀 Cooper makes constraints as easy to use as losses! Juan Ramirez Meraj Simon Lacoste-Julien Mila - Institut québécois d'IA 💻github.com/cooper-org/coo… 🧵👇