Observed arrival · 2026-09-21
Arshia Hemmat’s Uncertainty Lab
A Cambridge and Wellcome Sanger researcher’s portfolio on hallucination, uncertainty, trustworthy AI, and AI for biology.
- For
- Researchers and readers tracking trustworthy AI for biology
- Worth noticing
- The site describes a lab-in-the-loop process where model certainty helps select experiments that may consume a week of lab time.
Field notes
The homepage presents a research record spanning hallucination detection in diffusion language models, volumetric tissue spatial transcriptomics, abstract shape recognition, and local privacy for retrieval-augmented generation. Its central working premise is unusually concrete: when biological ground truth requires substantial laboratory time, uncertainty can influence which model-generated hypotheses receive experimental attention. The page also preserves a dated publication and affiliation timeline rather than presenting only a static biography.
Observed signals
Read the marks
Editorial observations of this landing page, not a rating.
One card from the complete issue