Devastatingly, we have lost a bright light in our field. Felix Hill was not only a deeply insightful thinker -- he was also a generous, thoughtful mentor to many researchers. He majorly changed my life, and I can't express how much I owe to him. Even now, Felix still has so much

The NREL community honors the legacy of former President Jimmy Carter. The lab owes a considerable debt to Carter. NREL was originally known as the Solar Energy Research Institute, which was founded by the Carter administration in 1977. bit.ly/3W2Sl74

#ICLR2025 is running the "Tiny Papers" track through the workshops this year! Great opportunity for researchers who are trad. underrepresented in ML & want to break into the community. There will also be financial assistance available for Tiny Paper authors. Learn more 👇

#ICML2025 includes a new track on Application-Driven Machine Learning (innovative ML techniques, problems, and datasets driven by the needs of end-users in real-world)! If this fits your work, consider submitting to ICML (dl: Jan 30) and checking the ADML box ✅ in OpenReview ⬇️

Severance is "2001: A Space Odyssey as an office" that's how production designer Jeremy Hindle described it. i pulled together these comparisons and it's true. enjoy.

"Regarding training usage: We acknowledge that OpenAI does have access to a large fraction of FrontierMath problems and solutions, with the exception of a unseen-by-OpenAI hold-out set ..." lesswrong.com/posts/cu2E8wgm… Lol. Lmao even

They're gonna roll back the most extreme stuff and everyone will breathe a sigh of relief, but ALL scientific funding and ALL foreign aid (except for Israel) is still on hold for who knows how long - a policy that a week ago would have been unimaginably extreme.

We are proud to announce that OpenAI has signed an agreement with the U.S. National Laboratories to supercharge their scientific research using our latest reasoning models. Sharing our technology with the nation’s top scientists aligns with our mission of building artificial

⏰Last chance to submit your work to our #ICLR2025 workshop. Submissions close at 23.59 AoE time today. We cannot wait to see all your exciting work!

Meet the recipients of the 2024 ACM A.M. Turing Award, Andrew G. Barto and Richard S. Sutton! They are recognized for developing the conceptual and algorithmic foundations of reinforcement learning. Please join us in congratulating the two recipients! bit.ly/4hpdsbD

ah, well, alas,,,

The lesson I’ve learned over the past seven years working on the Science of AI is this: (1) We can never fully understand how AI works. (2) But with a scientific mindset, we can always understand more. This perspective has fundamentally transformed how I see myself and others,

New article! On the effectiveness of neural operators at zero-shot weather downscaling 👉 bit.ly/4j9qB9W By Saumya Sinha, Brandon Benton and Patrick Emami (NREL) #MachineLearning #wind #downscaling

Can we use RL to plan with combinatorial constraints? Our #ICLR2025 paper combines deep RL with mathematical programming to do so! We embed a trained Q-network into a mixed-integer program, into which we can specify NP-hard constraints. w/ Bryan Wilder Elias Khalil Milind Tambe (moving @milindtambe-ai.bsky.social)

🗣️ If you still need lodging in Singapore for #ICLR2025, I have a hotel room reserved in the block (Conrad Singapore Orchard, April 24-29) that is now available. Willing to be ~flexible on cost if traveling on a budget. Send me a msg if interested!

can't really just use an LM as a surrogate for the world or a person. its internal causal model is pretty bespoke. GPT-2 and Llama have very different causal models of the world.

2025 NATIONAL CHAMPIONS!!!!!!!!!!!!!!!!!!!!

For every issue that goes viral, there are literally dozens that are dealt without much fanfare. In general, people grossly under-appreciate how much work Carl Vondrick and the other #ICLR2025 PCs do. Those attending Singapore in person should thank them personally. As a side

Test of Time Winner Adam: A Method for Stochastic Optimization Diederik P. Kingma, Jimmy Ba Adam revolutionized neural network training, enabling significantly faster convergence and more stable training across a wide variety of architectures and tasks.

We're excited to announce Yisong Yue and Iadine Chades as keynote speakers at our #ICLR2025 workshop on Apr 28. See you in Singapore! 🎉 Learn more about the workshop and our free livestream here: climatechange.ai/events/iclr2025 A big thanks to Mila - Institut québécois d'IA for sponsoring the event.