I am not at RL_Conference sadly! Regardless, I had an awesome time working as a sponsorship chair w/ brilliant @audurand ! Highly recommend the industry-academia mixer! My student Lynn Cherif , my colleagues from Mila - Institut québécois d'IA & Google DeepMind are there! Don’t be shy & say👋🏻🍒🥕

🚨 New paper alert! 🚨 A pipeline to generate increasingly difficult math evaluations using frontier LLMs, and human intervention! TLDR: Models show 📉 in performances on the generated questions ranging from 13.4% to 92.9%, relative to MATH, which bootstrapped the generation 🤯

Mila's annual supervision request process opens on October 15 to receive MSc and PhD applications for Fall 2025 admission! Join our community! More information here mila.quebec/en/prospective…

Stop by our poster on Parseval Regularization for Continual RL with the amazing Wes Chung, Doina Precup, and David Meger today 11 am - 2 pm West Ballroom poster 6304!

I am presenting the two following works (in🧵) done at Mila - Institut québécois d'IA on Sunday 3:00 PM at the SafeGenAI workshop NeurIPS Conference. Drop by and let’s chat! Papers:

ICLR 2025 many many many thanks to Kyunghyun Cho and Yoshua Bengio for enabling the wildest ever start of my research career 2014 was a very special time to do deep learning, a commit that changes 50 lines of code could give you a ToT award 10 years later 😲

Come check out our poster at #ICLR2025 at The Third Workshop in Deep Learning for Code Workshop (@DL4C)! ⏰ Mon, Apr 28 3pm-4:30pm – Garnet 218-219 I unfortunately won’t be there but my fantastic advisor @Khimya will be presenting our new paper – Cracking the Code of Action: a

Honored that our paper Temporal Difference Flows received the Best Paper Award at the #ICLR2025 World Models Workshop, and has also been accepted as a spotlight for #ICML2025! All made possible with the exceptional team AI at Meta! 📄arxiv.org/abs/2503.09817 x.com/JesseFarebro/s…

📢 If you're an RL researcher and want to build RL algorithms that are immediately applicable to the AI systems already being built by many thousands of programmers. Add your new LLMs reinforcement learning algorithm for agents/workflows to DSPy and Arbor.

Veronica Chelu from McGill University and Mila - Institut québécois d'IA shows how excitatory-inhibitory dynamics in recurrent neural networks can affect decision making and reinforcement learning. Veronica Chelu #NeuroAI2025 #ML #neuroscience

As AI agents face increasingly long and complex tasks, decomposing them into subtasks becomes increasingly appealing. But how do we discover such temporal structure? Hierarchical RL provides a natural formalism-yet many questions remain open. Here's our overview of the field🧵

Still in stealth but our team has grown to 20 and we're still hiring. If you're interested in joining the research frontier of deploying RL+LLM systems, shoot me an email to chat at ICML!