Our work in collaboration with the Gagliardi group: “Cartesian Equivariant Representations for Learning and Understanding Molecular Orbitals” accepted to the Proceedings of the National Academy of Sciences (PNAS) chemrxiv.org/engage/api-g... @danielgrzenda.bsky.social @nathaniel-hudson.bsky.social

We’re welcoming 5 new faculty members to our computer science department this semester—bringing expertise in AI, cybersecurity, HPC, cryptography, and more. https://bit.ly/46pyymr

📢 Exciting news! Our paper on Flight—a novel hierarchical federated learning framework built on Globus Compute, Parsl, and ProxyStore—is now officially published in Future Generation Computer Systems! ✈️🤖 Use the link below to check out the paper on FGCS: www.sciencedirect.com/science/arti...

It should have always been obvious that relying on LLMs to stochastically regurgitate text on behalf of users is catastrophic for learning outcomes. I hope to see continued research in this direction as more educational institutions choose to surrender to AI hype.

@mansisakarvadia.bsky.social, Aswathy (PhD students), and @nathaniel-hudson.bsky.social (postdoc) presented their work on identifying and ablating memorization in #LLMs at the 2024 MSLD workshop! 🎉 Their research is also accepted to ICLR 2025 — check it out: mansisak.com/memorization/

1/🧵ICLR 2025 Spotlight Research on LM & Memorization! Language models (LMs) often "memorize" data, leading to privacy risks. This paper explores ways to reduce that! Paper: arxiv.org/pdf/2410.02159 Code: github.com/msakarvadia/... Blog: mansisak.com/memorization/

Super proud to have had the opportunity to mentor Jordan on this project. He's bound for great things.

Our first ever post celebrates @globuslabs.bsky.social undergraduate intern Jordan Pettyjohn winning 1st place at the Student Research Competition at SC24 for his work "Mind Your Manners: Detoxifying Language Models" sc24.supercomputing.org/proceedings/...