Shanghua Gao
@GaoShanghua
Research Fellow at Harvard University
Excited to share our perspective paper in @CellCellPress on how AI agents can transform biomedical research! Unlike traditional one-step machine learning models, AI agents engage in multi-step, collaborative problem-solving.
Excited to share our perspective in @CellCellPress, where we discuss “AI scientists” as collaborative AI agents designed to empower biomedical research cell.com/cell/fulltext/… While the concept of an “AI scientist” is aspirational, advances in agent-based AI are paving the way…
🌍 Excited to open up our global evaluation of AI for drug decision-making and therapeutic reasoning @GaoShanghua Want to shape the future of therapeutic AI, from understanding existing medicines to developing new treatments for diseases with limited options? Start here:…
We provide an easy-to-use starter kit to help you join the competition effortlessly! Check here: github.com/mims-harvard/C…
Excited to see that 130+ teams have already entered CUREBench @NeurIPSConf 2025 Challenge Start here: curebench.ai Evaluate AI deeply and broadly for therapeutics @GaoShanghua @cziscience @MilkenInstitute @BiswasFamilyFdn @cziscience Rare as One program @harvardmed…
Join the CUREBench @NeurIPSConf 2025 Challenge at curebench.ai for therapeutic reasoning! We are excited to see that 130+ teams have joined! @cziscience @MilkenInstitute @BiswasFamilyFdn @cziscience Rare as One program @harvardmed @BrighamWomens @HarvardDBMI
📣 Competition launch alert! CUREBench competition at @NeurIPSConf 2025 Start here: curebench.ai Benchmarking AI for therapeutic reasoning, drug discovery, treatment planning, and therapeutic decision-making 🎯 Track 1: Develop AI models that rely on parametric…
Check out our ICML 25 work MedTok: A very powerful multimodal medical code tokenizer!
🚨 I'm excited to share that MedTok, our multimodal medical code tokenizer, has been accepted to #ICML2025! 🎉 Try out MedTok and see how it can enhance biomedical AI applications. huggingface.co/mims-harvard/M… Huge thanks to @marinkazitnik and all my amazing collaborators!…
Thrilled to launch CUREBench Competition @ #NeurIPS2025! Benchmark therapeutic reasoning & drug discovery with agentic judges & human experts. $40k prize pool + travel awards • Deadline Oct 15 2025 Join us ➡️ curebench.ai
📣 Competition launch alert! CUREBench competition at @NeurIPSConf 2025 Start here: curebench.ai Benchmarking AI for therapeutic reasoning, drug discovery, treatment planning, and therapeutic decision-making 🎯 Track 1: Develop AI models that rely on parametric…
Cures Within Reach @CuresWReach is now accepting submissions for Phase I or Phase IIA clinical trials to validate AI-driven drug repurposing opportunities in any unsolved disease! @ChanZuckerberg Eligible clinical trials must include AI-generated data as part of the preclinical…
Our new #ClinicalTrial funding RFP for trials in any disease based anywhere in the world to validate #AI-driven drug #repurposing is open until Aug 25! Info at bit.ly/cwrrfps. Thanks to our lead partner @ChanZuckerberg Help spread the word! #ClinicalTrials #RareDiseases
Introducing ATOMICA 💫 A model to universally represent molecular interactions (for proteins, nucleic acids, small molecules, and ions) at an all-atom scale 🧵
Big thanks to @KempnerInst for the incredible support — all our models are trained on the Kempner cluster!
NEW blog by #KempnerInstitute researchers @GaoShanghua & @marinkazitnik: introducing TxAgent, an #AI agent that redefines how AI can reason, retrieve & integrate biomedical knowledge for precision therapeutics. Read the blog: bit.ly/41VlrYc
Great to see TxAgent, our AI system for therapeutic reasoning, featured in the world’s largest AI newsletter!
Top stories in AI today: -China’s Baidu launches ultra-cheap AI models -How to code with AI, directly in your editor -Harvard team creates AI agent for medicine -4 new AI tools & 4 job opportunities Read more: therundown.ai/p/chinas-ultra…
👉 Preprint: arxiv.org/abs/2503.10970 👉 TxAgent: zitniklab.hms.harvard.edu/TxAgent 👉 Open code and data: github.com/mims-harvard/T… 👉 ToolUniverse: github.com/mims-harvard/T… We are excited about the next steps as we work with clinical researchers, disease foundations, and patient-led…
Thank you, Marinka, for introducing TxAgent and for your outstanding leadership on this project! We’re thrilled to see how TxAgent can help with therapeutics and how its techniques can be applied to other fields!
🚀 Introducing TxAgent: a first of its kind AI agent for therapeutic reasoning across a universe of 211 tools, with a comparison against DeepSeek-R1 671B @NVIDIAAI TxAgent is an AI agent that redefines how AI can reason, retrieve, and integrate biomedical knowledge for precision…
TxAgent An AI Agent for Therapeutic Reasoning Across a Universe of Tools
Work led in collaboration with #KempnerInstitute's @marinkazitnik is featured as a conference paper at #NeurIPS2024 today! Be sure to check out "UniTS: A Unified Multi-Task Time Series Model." @GaoShanghua @teddykoker @oq_35 @tom_hartvigsen buff.ly/41xUTgT
In the latest issue! Empowering biomedical discovery with AI agents dlvr.it/TG0lFM
Excited to share our perspective in @CellCellPress, where we discuss “AI scientists” as collaborative AI agents designed to empower biomedical research cell.com/cell/fulltext/… While the concept of an “AI scientist” is aspirational, advances in agent-based AI are paving the way…