Congratulations to Prof. Prakash Panangaden, Core Academic Member at Mila, on being elected a Fellow of the The Royal Society. This prestigious distinction recognizes the excellence of his work and his outstanding contributions to scientific research. royalsociety.org/news/2025/05/n…

The NeurIPS Position Track is seeking additional reviewers this year. If you'd like to serve or nominate a colleague, please complete our nomination form! docs.google.com/forms/d/e/1FAI…

Yesterday, we had the pleasure of announcing the seven recipients of Mila’s inaugural Sustainability Scholarships Program. Learn more about the laureates and their research projects: mila.quebec/en/news/seven-…

Want to shape AI policy for the public good? There’s still time to apply for Mila’s AI Policy Fellowship! Don't miss out - apply by June 2! ow.ly/quH550VWFko

Recent presentation on UI-Agents at Mila tea-talk. youtube.com/watch?v=oE5xHQ… Covers: * benchmarking * improving agent * agent security

anthropic alignment researcher tweeted this about opus and then deleted it. "contact the press" bro this company needs to be shut down now

Now accepted at Nature Communications. This should be the first study showing #plasticity effects in #brain #asymmetry in adults. Cross sectional data cannot show this at the same degree of evidence in our opinion .

In summary, while data can change the power law exponent significantly, it's quite challenging to find architectures and optimizers that change the exponent! If you want to see something that *does* change the power law exponent, stay tuned and follow Damien Ferbach for more 😁

The NeurIPS 2025 May Newsletter is now available on our blog: blog.neurips.cc/2025/05/22/neu… If you would like to receive these newsletters directly by email, please subscribe to the newsletter mailing list under your profile: neurips.cc/Profile/subscr…

Mila members David Rolnick, Yuyan Chen and Mélisande Teng have arrived in the tropical rainforest of Panama to test an algorithm to discover new species. Among a team of entomologists and computer scientists, they have installed camera traps to automatically monitor biodiversity.

Introducing our 3rd Keynote 🎤 Ghada Sokar, Research Scientist at Google DeepMind and Adjunct Professor at Tu/e, will speak on "Network Plasticity and Scalability in Deep Reinforcement Learning" — exploring model adaptation and continuous learning under non-stationary conditions

We’re excited to kick off the 2025 Summer School in Responsible AI and Human Rights with an incredible new cohort of participants! From insightful keynotes to hands-on workshops, the energy and curiosity in the room are already setting the tone for a transformative week. In

It's very difficult to improve the *exponent* in scaling laws for loss vs compute, especially by changing the optimizer! Our new paper shows that scaling momentum correctly can *provably* improve the scaling exponent on a theoretical model. Empirically, it works on LSTMs too!

Want to inform evidence-based policy on AI safety and governance? Apply by June 2 for Mila’s AI Policy Fellowship. This program brings together researchers, practitioners, and Mila experts to co-develop evidence-based recommendations for decision-makers. mila.quebec/en/ai4humanity…

The next ~1-4 years will be taking the 2017-2020 years of Deep RL and scaling up: exploration, generalization, long-horizon tasks, credit assignment, continual learning, multi-agent interaction! Lots of cool work to be done! 🎮🤖 But we shouldn't forget big lessons from back

Dear PhD students now regretting taking offers at US schools: If you turned down PhD offers in Canada, but want to rethink that, email the professors who were trying to recruit you. They might be able to pull some strings. Your sane neighbor to the north, Canada

Throughout the week, computer scientists and entomologists have collaborated to discover new species of moths in the tropical forest of Panama, an area with some of the most diverse biodiversity on the planet. By fine-tuning the Antenna platform together, Mila researchers and

Our remote sensing foundation model Galileo has been accepted to ICML 2025! Galileo outperforms state-of-the-art across different input data modalities and shapes, and using it requires only minimal data and compute. More at: arxiv.org/abs/2502.09356

Confused about recent LLM RL results where models improve without any ground-truth signal? We were too. Until we looked at the reported numbers of the Pre-RL models and realized they were serverely underreported across papers. We compiled discrepancies in a blog below🧵👇

I recently participated in the High-Level Expert Panel on AGI, convened by the Council of Presidents of the UN General Assembly, which has just released its final report on the Governance of the Transition to AGI. Report below ⬇️ uncpga.world/agi-uncpga-rep…