Faisal Mahmood
Built UNI and CONCH, the pathology foundation models that let AI read whole-slide images across cancer types.
Overview
Faisal Mahmood's laboratory developed CLAM for weakly supervised whole-slide learning, TOAD for predicting tumour origin, and the UNI and CONCH foundation models trained on more than 100 million pathology images, which set the benchmark for general-purpose computational pathology. His group also builds multimodal models integrating histology with genomics for prognosis and is a leading academic voice on AI in pathology.
| Title | Journal | Year |
|---|---|---|
| Towards a general-purpose foundation model for computational pathology (UNI) | Nature Medicine | 2024 |
| AI-based pathology predicts origins for cancers of unknown primary | Nature | 2021 |
| Data-efficient and weakly supervised computational pathology on whole-slide images | Nature Biomedical Engineering | 2021 |
Pages like this
not linked directly; found by shared links- PersonThomas J. Fuchs
Shares Pathology & radiology foundation models, Digital pathology & AI and the tags ai, pathology.
- PersonJakob Nikolas Kather
Shares Pathology & radiology foundation models, Digital pathology & AI and the tags ai, pathology.
- TechnologyVirchow / Virchow2 (Paige, MSK)
Shares Pathology & radiology foundation models, Digital pathology & AI and the tag pathology.
- TechnologyCHIEF (Harvard, Yu Lab)
Shares Pathology & radiology foundation models and the tag pathology.
- TechnologyPhikon / Phikon-v2 (Owkin)
Shares Pathology & radiology foundation models and the tag pathology.
- TechnologyPLUTO (PathAI)
Shares Pathology & radiology foundation models and the tag pathology.
- IdeaContinuous prospective validation for every oncology AI tool after deployment
Shares Pathology & radiology foundation models, Digital pathology & AI.
- TechnologyUNI and CONCH (Harvard, Mahmood Lab)
Shares Pathology & radiology foundation models and the tag pathology.