Tag: explainable AI
A Singapore vision model reads 18 datasets and points at the evidence
Researchers at A*STAR and Singapore Eye Research Institute built a medical vision language model trained on 1.3 million samples across…
A single mistyped probability nearly broke a transparent cancer model
Leipzig's Bayesian network matched tumor board decisions 95.1% of the time, and an audit of its own tables explains why.
Deciphex’s imaging engine traces every finding back to the tissue
The Irish company says its CipherX system names the structures on a slide that produced a prediction.
Explainable AI model ConceptCLIP shows clinicians its reasoning
ConceptCLIP pairs strong diagnostic accuracy with explanations clinicians can verify, a step toward trustworthy medical AI.
Skin AI explanations help doctors while biasing everyday users
New Nature Medicine experiments with 623 lay people and 153 physicians show explainable AI boosts experts but fuels automation bias…