If you like #AI and LEGO you will like this. We have implemented an interface to train LEGO robots in #PyTorchRL from PyTorch using Pybricks . Examples, codes, etc all available. Collaboration with AI at Meta. bricksrl.github.io/ProjectPage/ arxiv.org/abs/2406.17490

This is probably one of the best opportunities for a junior bio researcher right now who wants to learn how the wet lab and AI will interact in the future of AI-powered science. If that sounds like you, get in touch!

We are already on Day 3 of the CECAM workshop "L2M3: Large language models for materials, molecules and beyond" in CECAM-HQ (EPFL) ☀️ A fascinating panel discussion on "Agents for AI 4 science" with Santiago Miret, Tong Xie, Philippe Schwaller (he/him) and Timur Madzhidov is currently underway.

This has been a great workshop! Thanks CECAM and Kevin M Jablonka for putting this together!

And if you want to build things with related techniques, we have various openings for PhDs, but now also with FAIRmat fairmat-nfdi.eu/jobs/Generativ…

Join us at PolymathicAI at Simons Foundation's Flatiron Institute to create AI models for science! Opportunities available for postdoctoral candidates! Google form below: forms.gle/JbdWA3VczH3Xfg…

LAB-Bench is, to my knowledge, the first set of evals designed to measure whether models/agents can do scientific research, not just whether they know assorted trivia about science. Procedural evaluations like this for complex tasks will be very important going forward.

The team that pulled together LAB-Bench, led by Jon Laurent and Joseph D. Janizek, is extremely cracked. They pulled this together, from beginning to end, in just over 2 months. Lots more to do. If you want to join: job-boards.greenhouse.io/futurehouse/jo… Major contributions also from Michael

Join us! We need exceptionally talented bioengineers, ML researchers, and everyone in between to help us build and evaluate AI systems to automate science.

Excited about our just-out Nature article! 🌟 A standout: a machine learning study where we used net-carbon avoidance cost to identify top-performing materials. Started with 20, #MachineLearning added 135 more, boosting commercialization chances! doi.org/10.1038/s41586…

If you want to extract data using LLMs, you should check out our review 📝 arXiv arxiv.org/abs/2407.16867 💻 hands-on online book matextract.pub

Kevin included an entire interactive book with his review - including an example of multi-modal LLMs to automatically assess NMR spectra in figures! This is a tutorial in the book. This is a great resource for those wanting to apply LLMs/agents to chemistry and materials

Nawaf Martiño Ríos-García Mara Schilling-Wilhelmi Adrian Mirza is at Vienna. We are about to spam #AI4Mat with talks and posters. - 💡 two spotlight talks - 🖼️ four posters 🔥🚀 If you are at #AI4Mat catchup with us. #Matsci #BOKU #ICML #Materials

Friends, I am beyond happy! I'm starting a new position as Full Professor of #MachineLearning at the University of Fribourg Uni Fribourg 🇨🇭! With #SwissAI and many other initiatives, I am taking my research at the intersection of #geometry, #topology, and #MachineLearning to a new

in Gastown with Kevin M Jablonka and Mohamad Moosavi during the Accelerate Conference!

Rigorous methods sections alone would go a long way to improving the scientific enterprise, with an LLM that reads your paper draft & figures and tells the authors/journal what additional details they should be including

Introducing PaperQA2, the first AI agent that conducts entire scientific literature reviews on its own. PaperQA2 is also the first agent to beat PhD and Postdoc-level biology researchers on multiple literature research tasks, as measured both by accuracy on objective benchmarks