A new Perspective in Cell, co-authored by HIP Co-Chief Science Officer John Tsang alongside leading researchers from Columbia, Stanford, and the Chan Zuckerberg Initiative, tackles one of the biggest open questions in biomedical AI: why hasn’t generative AI’s success in protein folding translated to predicting how cells and tissues actually behave? Drawing inspiration from Hilbert’s landmark list of 23 mathematical problems, the authors propose fifteen concrete challenges to focus the field — spanning molecular interactions, cell-state reprogramming, drug response, and clinical trial prediction — each paired with rigorous benchmarks for measuring real progress.
Spoiler alert: HIP is called out as a model for what comes next.
Fifteen challenges for generative AI applications to cell biology
Responses