Xiaorui Su
@xiaorui_su
Postdoc Fellow in Harvard Medical School @marinkazitnik @harvardmed
🌍 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:…
Data is everything. We need more data to unlock the next scientific breakthroughs with AI. Here’s how we’re addressing this at NeurIPS. With a $10,000 prize pool to get it started. 🧵 @NeurIPSConf @AI_for_Science @RenPhil21 #NeurIPS2025
'Multimodal Medical Code Tokenizer' @xiaorui_su · @shvat_messica · @YepHuang · Ruth Johnson · Lukas Fesser · @GaoShanghua · Faryad Sahneh · @marinkazitnik doi.org/10.48550/arXiv… (7/20)
📣 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…
Join us at ICML @icmlconf this Thu 17 Jul from 4:30-7 pm PT to learn about MedTok and try it out! Congrats to @xiaorui_su @shvat_messica MedTok is a general-purpose tokenizer of medical codes that can be integrated into any transformer or tokenization based biomedical AI By…
New paper: MedTok – A general multimodal medical Code Tokenizer @xiaorui_su MedTok is a drop-in replacement for standard tokenizers, improving structured data representation in clinical prediction models, medical QA systems, and medical foundation models EHR foundation models…
📢 One patient, many contexts, yet our AIs are still refreshing outdated prompts We envision context-switching AI will adapt to infinitely many medical contexts, new medical specialties, healthcare roles, diseases, and populations @_michellemli arxiv.org/abs/2506.10157 ♾…
TxAgent is an agent that integrates 211 tools for therapeutic evidence based reasoning. I am so thrilled to be involved in this fantastic and amazing project!
🚀 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…
Predicting clinical outcomes of drug combinations from preclinical data is a major challenge @YepHuang We know a drug works in the lab. But will it work in patients? 🔬 ➡️ 🏥 This is key for safe and effective therapies and it's one of the hardest challenges in medicine.…
New paper: MedTok – A general multimodal medical Code Tokenizer @xiaorui_su MedTok is a drop-in replacement for standard tokenizers, improving structured data representation in clinical prediction models, medical QA systems, and medical foundation models EHR foundation models…
Introducing KGARevion, a KG+LLM agent 🤖designed for knowledge-intensive medical QA. Led by @xiaorui_su @GaoShanghua @valegiunca Here's why it is unique: 1️⃣ Open-ended reasoning: Generates nuanced answers without relying on predefined options 2️⃣ Robustness: LLMs can be…
Our @marinkazitnik lab team at the @HarvardCMSA 2024 Ding Shum lecture by @ylecun today. We missed getting a photo with the brilliant speaker himself – but maybe that’s another incentive to invite @ylecun back to Harvard again? 😁 1/4
UniTS: Building a Unified Time Series Model - Supports a universal task specification, accommodating classification, forecasting, etc - Remarkable zero/few-shot and prompt learning capabilities when evaluated on new data domains and tasks repo: github.com/mims-harvard/U… abs:…
📝PocketGen: a deep generative method for generating the residue sequence and the full-atom structure within the protein pocket region that leverages sequence-structure consistency. biorxiv.org/content/10.110…