Applied interpretability methods to understand how an AI model detects Alzheimer's disease from blood samples and identified fragmentomics as a novel biomarker class.
How media typically covers Nicholas Wang
Referenced in coverage
Using interpretability methods on an AI model trained to detect Alzheimer's from blood samples, researchers identified DNA fragment length patterns (fragmentomics) as a novel biomarker class that generalizes better than previously reported biomarkers when tested on independent cohorts.
“Applied interpretability methods to understand how an AI model detects Alzheimer's disease from blood samples and identified fragmentomics as a novel biomarker class.”