#clinical-ml News & Analysis

3 articles tagged with #clinical-ml. AI-curated summaries with sentiment analysis and key takeaways from 50+ sources.

3 articles

AINeutralarXiv – CS AI · Jun 116/10

🧠

Using Explainability as a Training-Time Reliability Signal for Efficient ECG Classification

Researchers introduce ERTS, an explainability-based training method that reduces computational costs for ECG classification by using attention map quality to identify which training samples are genuinely informative versus noisy. The approach demonstrates consistent performance improvements across multiple datasets while significantly lowering training expenses, offering practical efficiency gains for resource-constrained healthcare environments.

AINeutralarXiv – CS AI · Jun 26/10

🧠

Motif-based morphology signatures for interpretable ECG screening and monitoring

Researchers propose a motif-based framework for ECG analysis that identifies interpretable cardiac signatures through beat-aligned morphology patterns, enabling early detection of cardiovascular abnormalities. Using Dynamic Time Warping to extract representative cardiac cycles, the method quantifies morphological drift across short and long-term monitoring with three metrics: deviation from normal sinus rhythm, personalized baseline deviation, and motif instability. Testing on standard ECG datasets demonstrates significant separation between normal and arrhythmic subjects with high statistical significance.

AINeutralarXiv – CS AI · May 285/10

🧠

GraD-IBD: Graph Representation Learning from Diagnosis Trajectories for Early Detection of Inflammatory Bowel Disease

Researchers propose GraD-IBD, a graph-based machine learning model that analyzes patient diagnosis histories encoded in ICD codes to detect inflammatory bowel disease risk earlier and more efficiently than existing sequential models. The approach reformulates longitudinal diagnostic trajectories as temporally directed graphs with a novel message-passing mechanism, demonstrating improved accuracy while reducing computational complexity.