Overview
- South Ural State University reported PADDi, a PALMAD-based anomaly detection algorithm that runs across distributed GPUs and was tested on the Lomonosov-2 and Lobachevsky supercomputers.
- PADDi divides massive ECG time series into fragments and segments for parallel processing and uses a formalized concept of dissonance to flag anomalies without relying on domain specialists.
- The Ministry of Science announced that Penza State University and Penza State Technological University patented a neural-network method that analyzes ECGs together with fluorography.
- The Penza approach accounts for both the electrical and geometric axes of the heart, targets detection of severe conditions, and produces a preliminary AI conclusion that a clinician reviews.
- Developers say the Penza system was trained on data from more than 250 patients and is slated for deployment in local clinics and hospital wards, while broader clinical validation and scaling remain ahead.