Harvard team proposes biomedical world models framework in Cell

Marinka Zitnik's team at Harvard Medical School and the Kempner Institute has published "World models for biomedicine" in Cell's "AI in Biology" special issue. The paper brings the AI concept of world models—systems that represent states and internally simulate their evolution before taking action—into biomedicine, spanning multiscale applications from molecules to patients. Led by marinkazitnik's team (per a repost, ayushnoori and njwfish co-led the work). The takeaway: biology needs models that don't just describe the world but predict what happens after an intervention—simulating the consequences of interventions before running actual experiments.

Confirmed

• The paper "World models for biomedicine" was published in Cell's "AI in Biology" special issue, with authors from Harvard Medical School, the Kempner Institute, and other institutions
• Core claim: biomedical world models should predict intervention outcomes rather than merely describe the world
• The paper systematically lays out the data types, training approaches, and evaluation methods needed to build biomedical world models, and is regarded as a framework reference for researchers at the AI–biology intersection

Why it matters

• The authors identify three key challenges: biological states are typically only partially observable; destructive measurements limit repeated measurement of the same system; and intervention data is scarce, making it hard to learn causal relationships between interventions and outcomes
• If world models can simulate the consequences of interventions before experiments, they could substantially reduce the cost and risk of biomedical experiments and accelerate translation from basic research to clinical applications

2026-09-18 ~ 2026-09-18 · 8 related posts

• Cell Focus: Biology Needs World Models That Predict Interventions, Not Just Descriptions — rishabh16_ · 2026-09-18
• [source] Cell paper lays out how world models could transform biomedicine, from molecules to patients — marinkazitnik · 2026-09-18
• [source] Biomedical world models: a framework for virtual drug trials, intervention design and planning — marinkazitnik · 2026-09-18
• [source] What makes a world model? A four-part test that foundation models, digital twins fail — marinkazitnik · 2026-09-18
• Three Open Challenges for Biomedical World Models, Per Marinka Zitnik — marinkazitnik · 2026-09-18
• New Cell Paper: World Models to Simulate Biomedical Interventions Before Running Them — marinkazitnik · 2026-09-18
• Harvard team's Cell paper lays out three data challenges for biomedical world models — marinkazitnik · 2026-09-18

1 near-duplicate retellings: marinkazitnik