Thrilled to release 🌟STaRK 🌟 - A large-scale LLM retrieval benchmark on semi-structured knowledge bases.

While LLMs excel at reasoning and semantic retrieval, they struggle with more complex tasks. Especially when real-world user queries require a combination of unstructured…

Thank you Stanford University for highlighting our work on #genAI for designing new molecules and drugs! med.stanford.edu/news/all-news/…

Sharing a blog post I wrote about our recent work on ADMET-AI for fast and accurate prediction of important druglike properties of molecules. Check it out below and try ADMET-AI at admet.ai.greenstonebio.com

Here is the non-paywall link for our #SyntheMol paper rdcu.be/dB9z1

Researchers invent artificial intelligence model to design new superbug-fighting antibiotics #McMaster #Stanford brighterworld.mcmaster.ca/articles/artif…

If you're interested in ADMET prediction for drug discovery, please check out ADMET-AI, a free, open-source AI platform for large-scale ADMET prediction that I built with Greenstone Biosciences and James Zou. See the thread for more details and try it out at admet.ai.greenstonebio.com!

Kyle Swanson and I started down a dangerous road (argparse) and ended up on a beautiful mountain (Tap).

We're excited to celebrate 1 million downloads of Tap, a typed argument parser! We are grateful to the many contributors and users.

github.com/swansonk14/typ…

I really enjoyed talking with Jaanak and Jaagat about AI for drug discovery and biomedicine on their excellent STARTS Podcast about STEM and the arts. You can check out the episode below!

Method alert: Machine learning can help with the challenge of interpreting imaging of 40+ protein biomarkers enabled by CODEX systems. A machine learning framework called 7-UP can classify cell types and predict patient survival outcomes. In PNAS Nexus: ow.ly/U5FJ50OBKKH

An interdisciplinary team of remarkable scientists (and great friends) put together a study leveraging AI to find new antibacterial molecules against Acinetobacter baumannii. I think we are well into an era of AI-augmented drug discovery...WOOO!!!

nature.com/articles/s4158…