Read about AI in medicine in the latest Harvard Gazette and New York Times.

TxAgent is an AI agent for therapeutic reasoning that consolidates 211 tools from trusted sources, including all US FDA-approved drugs since 1939 and validated clinical insights. [Project website] [TxAgent] [ToolUniverse]

New paper: Multimodal AI approach for designing combination therapies with improved predictive accuracy and clinical relevance. [Project website]

KGARevion is an AI agent designed for complex biomedical QA that integrates the non-codified knowledge of LLMs with the structured, codified knowledge found in knowledge graphs. [ICLR 2025 publication]

Meet MedTok, a multimodal medical code tokenizer that transforms how AI understands structured medical data. By integrating textual descriptions and relational contexts, MedTok enhances tokenization for transformer-based models—powering everything from EHR foundation models to medical QA. [Project website]

What if we could anticipate molecular and medical changes before they happen? Introducing CLEF, an approach for counterfactual generation in biological and medical sequence models. [Project website]

New paper on the role of digital twins as global health and disease learning models for preventive and personalized medicine.

New papers on test-time interventions in language models and knowledge graph based LLM agents accepted to ICLR. [KGARevion]

Excited to share our new graduate course on Artificial Intelligence in Medicine 2.

Thanks to Kempner Institute for highlighting our latest research, ProCyon, our protein-text foundation model for modeling protein functions.

New Voices piece in Cell Systems: How will computational protein design change biotechnology and therapeutic development?

New paper: ProCyon is a groundbreaking foundation model for modeling, generating, and predicting protein phenotypes. [Project website] [Code]

New paper: A unified resource provides a new representation of clinical knowledge by unifying medical vocabularies. (1) Phenotype risk score analysis across 4.57 million patients, (2) Inter-institutional clinician panels evaluate alignment with clinical knowledge across 90 diseases and 3,000 clinical codes.

Are biomedical AI models truly as smart as they seem? SPECTRA is a framework that evaluates models by considering the full spectrum of cross-split overlap: train-test similarity. SPECTRA reveals gaps in benchmarks for molecular sequence data across 19 models, including LLMs, GNNs, diffusion models, and conv nets.

Congratulations to Ayush Noori on being named a Rhodes Scholar! Such an incredible achievement!

PocketGen is a multimodal sequence-structure generative model for designing full-atom ligand-binding protein pockets.

We envision “AI scientists” as AI agents capable of skeptical learning and reasoning that empower biomedical research by integrating ML models and biomedical tools with experimental platforms.

New paper on activity-cliff informed contrastive learning for molecular property prediction.

New paper introducing a knowledge graph agent for complex, knowledge-intensive medical reasoning.

Exciting projects include a unified multi-task time series model, a flow-matching approach for generating protein pockets using geometric priors, and a tokenization method that produces invariant molecular representations for integration into large language models.