🚀 Today, we’re excited to introduce SimpleTES for scaling the scientific discovery loop. 🧵 I always ask myself: what are we actually scaling in scientific discovery? Most LLM discovery methods focus on test-time scaling generation — more tokens, more agents, more turns. But science advances through the evaluation-driven loops: propose → evaluate → refine → repeat. SimleTES captures this idea, discovering SOTA solutions across 21 scientific problems! Key discoveries: 🏎️ 2.17x faster lasso solver than glmnet — the gold-standard LASSO solver, engineered for decades. ⚛️ 24.5% fewer quantum routing overhead on IBM Q20 — superior than previous standard library LightSABRE. 📐 0.380868 on Erdős Minimum Overlap — outperforming previous solutions from mixed-frontier ensembles or humans. 🧬 0.74 on Tabula Muris (scRNA-seq denoising) — new SOTA, generalizing to unseen tissue types without retraining. #LLM #AI4Science #ScalingLaws #SimpleTES #MachineLearning

In AI for scientific discovery, the bottleneck isn't always generation — it's quite often evaluation. How do you design evaluators close to gold? Prevent reward hacking? And critically, how do you scale the evaluation-driven loop to reach genuinely novel discoveries?

Extremely excited about the results of @adaptyvbio RBX1 binder design competition! 𝑩𝒊𝒏𝒅𝑪𝒓𝒂𝒇𝒕2 performed very well, with 3 out of 7 designs binding to the disordered tail. Overall, only 9 binders worked out of 322 tested, 2.8% hit rate! Proud of the BC2 team ♥️

SyntheMol-RL has now been published! SyntheMol-RL is a reinforcement learning model for synthesizable small molecule drug design. We used it to design antibiotic candidates for the bacteria S. aureus with hits validated in vitro and in vivo in mice. 1/6 link.springer.com/article/10…

Finally getting to share one of my favorite projects. ICLR Oral! 🏆 It’s so strange how rigid video tokenization is. Think about it: why should a still landscape cost the same amount of tokens as a busy street? We built InfoTok. We went back to basics with Shannon’s information theory to make tokens "adaptive" in a principled way. Its 2.3x better compression and 11x faster inference demonstrates the magic of the old-school theory ✨ Check it out: research.nvidia.com/labs/dir…

Excited to share that our paper has been published in Nature Machine Intelligence! We conducted a randomized controlled trial at ICLR 2025 with 20,000 reviews to test whether LLM feedback improves peer review quality. Link: nature.com/articles/s42256-0…

To make a long story short, we uncover dozens of regions of our genome that control whether the virus persists or is cleared quickly. Further, we show that persistent EBV may serve as a biomarker of complex diseases-- from respiratory disease to autoimmunity.

🤔Want a principled way to RL your diffusion model? Check Data-regularized Reinforcement Learning (DDRL)! Post-train @nvidia #Cosmos World Foundation models with a million GPU hours! 🤯 Novel formulation ➡️ Theoretically integrates SFT into RL ➡️ Robust to Reward Hacking 🛑 Details: research.nvidia.com/labs/dir… #DDRL #Diffusion #RL #NVIDIA #Cosmos

Super impressed that, when @ElanaPearl wasn't happy with the loss curve, she realized she needed a PyTorch PR to fix it. A great read.

Congrats to @ElanaPearl for her awesome interPLM paper, now in Nature methods. A great way to explore the inner workings of protein language models, with a very well organized and easy-to-use codebase!

📢 Excited that #unitox is selected as a #NeurIPS2024 spotlight!💡 We created #LLM agent to analyze >100K pages of FDA docs from all approved drug ➡️ new database annotating 8 toxicity types for 2400 drugs. Validated by clinicians. openreview.net/pdf?id=Vb1vVr… Data zou-group.github.io/UniTox-w… Great job led by @JakeSilberg @KyleWSwanson @ElanaPearl! Thanks to Angela Zhang and @xaniarg for clinical expertise wonderful @genmab collaborators 👏

Congratulations to our best submission award winners!! 🏆 “Can Large Language Models Explain Their Internal Mechanisms?” by @nadamused_, @ghandeharioun, @RyanMullins, @emilyrreif, Jimbo Wilson, @Nithum, and @iislucas 🏆 “The Illustrated AlphaFold” @ElanaPearl and @JakeSilberg

First up, watch @ElanaPearl and @JakeSilberg present “The Illustrated AlphaFold” 🧬elanapearl.github.io/blog/20…

Share your best resources to learn about AlphaFold in the comments! This is one of the best blog posts to learn about AlphaFold that I've seen (by @ElanaPearl & @JakeSilberg): elanapearl.github.io/blog/20…

The Illustrated AlphaFold bit.ly/the-illustrated-af3 Do you want to know how AlphaFold3 works? It has one of the most intimidating transformer-based architectures, so to make it approachable, we made a visual walkthrough inspired by @JayAlammar's Illustrated Transformer! 🧵 (1/7)

Solidarity with Ukraine ✊ (Russian Embassy, London)

We went from being mad to making change happen. 40 laws signed to increase housing production and access this legislative session alone in California. gov.ca.gov/2022/09/28/califo…

Next time you dip your asparagus in salsa, remember the hands that harvested those tomatoes. ❤️ #WeFeedYou

This is a day I’ve dreamed of my whole life, this is the reason @DeepMind was founded, to build AI and use it accomplish extraordinary scientific breakthroughs like #AlphaFold 2, to advance science and benefit humanity. I could not be more proud of the incredible team!