Huge congratulations to @DemisHassabis and John Jumper on being awarded the 2024 Nobel Prize in Chemistry for protein structure prediction with #AlphaFold, along with David Baker for computational protein design. This is a monumental achievement for AI, for computational

Excited to share our new paper on "Diffusion Model Predictive Control" (D-MPC). Key idea: leverage diffusion models to learn a trajectory-level (not just single-step) world model to mitigate compounding errors when doing rollouts. arxiv.org/abs/2410.05364

Excited to share our new paper on "Diffusion Model Predictive Control" (D-MPC). Key idea: leverage diffusion models to learn a trajectory-level (not just single-step) world model to mitigate compounding errors when doing rollouts. arxiv.org/abs/2410.05364

Gemini 2.0 Flash is here! 🎊 We're quite excited about this model. It's better than Gemini 1.5 Pro on most benchmarks, but comparable in speed and latency to Gemini 1.5 Flash. It has multilingual native audio output, native image output, native tool use, and a new multimodal

It will also be extremely important to analyze the strengths and limitations of the new system. Here are some examples of tasks that o3 couldn't solve on high-compute settings (even as it was generating millions of CoT search tokens and consuming thousands of dollars of compute

Fun facts: federal grants also fund research on cardiovascular, infectious and chronic respiratory diseases, cancer, diabetes, Alzheimers, Parkinsons, epilepsy... these account for > 90% of human death globally. For reference: homicide < 1% ; war~0.2%; terrorism~0.05%

Making LLMs run efficiently can feel scary, but scaling isn’t magic, it’s math! We wanted to demystify the “systems view” of LLMs and wrote a little textbook called “How To Scale Your Model” which we’re releasing today. 1/n

Happy to share our new preprint “Improving Transformer World Models for Data-Efficient RL”: arxiv.org/abs/2502.01591 We propose a ladder of improvements to model-based RL and achieve for the first time a superhuman reward on the challenging Craftax-classic benchmark! 1/10

Better diffusions with scoring rules! Fewer, larger denoising steps using distributional losses; learn the posterior distribution of clean samples given the noisy versions. arxiv.org/pdf/2502.02483 Alexandre Galashov Valentin De Bortoli Guntupalli Guangyao (Stannis) Zhou Kevin Patrick Murphy Arnaud Doucet

Diffusion models are trained to predict the mean E[X_0 | X_t] of clean data given a noisy sample. We propose a novel method of learning the full posterior distribution p(X_0 | X_t) using scoring rules, which improves performance of few-step diffusion models. Check our paper -

Happy to have this published ...finally. Have a look if you are interested in understanding cortical columns, microcircuits, and the thalamus. Stay tuned for threads unpacking the paper... Miguel Lázaro-Gredilla antoine dedieu Joseph Marino Science Advances science.org/doi/10.1126/sc…

Can LLMs be used to discover interpretable models of human and animal behavior?🤔 Turns out: yes! Thrilled to share our latest preprint where we used FunSearch to automatically discover symbolic cognitive models of behavior. 1/12

1/12 How do animals build an internal map of the world? In our new paper, we tracked thousands of neurons in mouse CA1 over days/weeks as they learned a VR navigation task. Nelson Spruston HHMI | Janelia, w/ co-1st author Johan Winnubst Video summary: youtube.com/watch?v=yw_4uV… Paper:

Exciting! 'Space is a latent sequence' & CSCG theory of hippocampus we developed at @GoogleDeepmind and Vicarious gets strong support in this nature paper with a set of ingenious animal experiments, neural recordings and analysis. Congrats Weinan Sun Nelson Spruston Johan Winnubst...1

1/ Gemini 2.5 is here, and it’s our most intelligent AI model ever. Our first 2.5 model, Gemini 2.5 Pro Experimental is a state-of-the-art thinking model, leading in a wide range of benchmarks – with impressive improvements in enhanced reasoning and coding and now #1 on

Think you know Gemini? 🤔 Think again. Meet Gemini 2.5: our most intelligent model 💡 The first release is Pro Experimental, which is state-of-the-art across many benchmarks - meaning it can handle complex problems and give more accurate responses. Try it now →

🥁Introducing Gemini 2.5, our most intelligent model with impressive capabilities in advanced reasoning and coding. Now integrating thinking capabilities, 2.5 Pro Experimental is our most performant Gemini model yet. It’s #1 on lmarena.ai (formerly lmsys.org) leaderboard. 🥇

Gemini 2.5 Pro is an awesome state-of-the-art model, no.1 on LMArena by a whopping +39 ELO points, with significant improvements across the board in multimodal reasoning, coding & STEM. You can try it out now in AI Studio ai.dev & Google Gemini App with Gemini Advanced

Bars! #veo3

Prompt Theory (Made with Veo 3) What if AI-generated characters refused to believe they were AI-generated?