Congratulations to the SUPERB Team! Our work on the Speech Processing Universal PERformance Benchmark (SUPERB) has been ranked 7th among the most cited papers at INTERSPEECH over the past five years! A big round of applause to everyone involved.

Exciting work lead by Sherry Yang exploring hierarchical models (LLMs+diffusion) for material search given a language query/specification. See thread 👇for details

This was a really fun collaboration between experts in materials science, diffusion, and foundation models. Connecting material generation with foundation models (although still loosely in this work) has so many exciting implications!

Wow. Google is cleaning up at the Nobels this year. Yesterday's prize recognized the old Google Brain and today's goes to Google DeepMind! The torch has been passed from Bell Labs.

Excited to share our new paper “Efficient Exploratory Synthesis of Quaternary Cesium Chlorides Guided by In Silico Predictions” in J. Am. Chem. Soc.. Exploratory synthesis is expensive and target prioritization is critical to improve chances of synthesis. pubs.acs.org/doi/full/10.10…

Cluster expansion-driven Monte Carlo simulations revealed the temperature dependence of Li/M order/disorder, corroborating experimental findings.

Overall, this approach led to the successful synthesis and characterization of new polymorphs (e.g. for Cs2LiCrCl6 and Cs2LiRuCl6), and a new compound Cs2LiIrCl6, demonstrating the effectiveness of combining computational and experimental techniques in solid-state synthesis.

Really cool work from our materials and chemistry team Google DeepMind showing how guidance from GNoME, our AI model predicting computationally stable materials, can accelerate exploratory synthesis of an important class of materials in the lab.

Interested in working on materials science and AI at Google DeepMind? We’re hiring a research engineer, details here: boards.greenhouse.io/deepmind/jobs/…

Google DeepMind's public policy team have published an article that describes the philosophy and practical experience of the Google DeepMind Science team and how we collaborate with our various external experts and partners.

Waymo's market share is now equal to Lyft within SF. Incredible. Network effects is one of the best sources of defensibility. But it's proven to be not that important in ridesharing. You need a minimum network size, but once you have that, there are diminishing returns. In

📢Muratahan Aykol, Ekin Dogus Cubuk and colleagues from Google DeepMind introduce a computational approach to predict the most likely crystallization products from amorphous precursors, which has the potential to help with the synthesis of new materials. nature.com/articles/s4358…

Our paper on predicting the emergence of crystals from amorphous precursors with deep learning potentials is now published in Nature Computational Science! 🎉 Google DeepMind

Congrats to Ekin Dogus Cubuk, Muratahan Aykol, and the entire Google DeepMind team on a very interesting publication. The momentum is continuing to build in this field - the future of science is here 🔥

Predicting metastable crystals from amorphous precursors with deep learning Identifying which crystalline phases emerge first from an amorphous material is vital for advancing technologies in ceramics, electronics, and energy storage. Yet this has remained a challenge due to the

Muratahan Aykol Ekin Dogus Cubuk Google DeepMind An accompanying News & Views for this paper by J. C. Schön is now available! nature.com/articles/s4358… 🔓rdcu.be/d9CXZ

A very nice perpective on our recent approach to predicting crystal structures from amorphous precursors using deep learning potentials, by Prof. Schön.

Running the a2c workflow with MACE-MPA-0 + Cell relaxation Animation shows the steps - Soft sphere relaxation - Melt-Quench MD with MACE - Full relaxation for the subcell structures The last frames show that the method recovers the Si-dc structure to be the lowest energy

x.com/i/article/1895…

As AI capability continues to improve and becomes ubiquitous, a differentiator of products will come from effectively making contact with their industry and solving their specific problems. Congrats Mirror Mirror for doing this and nailing fashion aesthetics in image generation!