The 2024 #NobelPrize laureates in chemistry Demis Hassabis and John Jumper have successfully utilised artificial intelligence to predict the structure of almost all known proteins. In 2020, Hassabis and Jumper presented an AI model called AlphaFold2. With its help, they have

My student Zhishuai Liu will be presenting this paper on #NeurIPS2024 Poster ID: 6404 Room: West ballroom A-D Time: 4:30-7:30 pacific time, Dec 11 2024. Come to chat with us if you are interested in robust RL, sim-to-real gap, offline RL.

Our students Weixin @Kamome_233 and Hao-Lin Hao-Lun Hsu will be presenting this paper at #NeurIPS2024 today from 4:30pm to 7:30pm! Room: West Ballroom A-D Poster ID: 6501 Come to chat with us on how to generalize Thompson sampling to general reward function and noise setting.

So many layers of irony here ……

Dr. Ed Hammond, a pioneer in bioinformatics, has significantly advanced the field through his work on computer-based medical records and health information exchange standards. Duke will honor his contributions with a symposium on Jan. 9-10, 2025. biostat.duke.edu/news/duke-symp…

The most interesting bits of information: - no PRM or step-by step-verifier needed - PPO on {question, answer_i} pairs; using an advantage function based on accuracy of final answer and format. - RL-tuning a small model is WORSE than distilling from larger RL tuned model!

If you’re using the #ICML LaTeX template, there’s a typo in algorithmic.sty that prevents cross-referencing specific lines in the algorithm environment. The fix is simple: change \addtocounter{ALC@line}{1} to \refstepcounter{ALC@line} on Line 106. Credit: tex.stackexchange.com/questions/5234…

Many academia labs (including mine) couldn't even afford a single H100 server. There are much better ways to spend money than 500B mostly on GPUs for one company, which already has plenty of them. That money could lead to extraordinary innovation in academia.

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

Just realized your #ICML2025 reviewers may have used Official Comment instead of Rebuttal Comment to reply to authors's rebuttal, and the former is invisible to authors so far. I've alerted reviewers in my AC batch to resend. I wonder how widespread this mistake is?

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🧵👇

Perfectly said. The US's dominance in science and technology is now at risk.

Similar observation: I've never seen reviews for my batch of ACing papers submitted so quickly.

Official results are in - Gemini achieved gold-medal level in the International Mathematical Olympiad! 🏆 An advanced version was able to solve 5 out of 6 problems. Incredible progress - huge congrats to Thang Luong and the team! deepmind.google/discover/blog/…