A 43.6 TOPS/W AI Classifier with Sensor Fusion for Sepsis Onset Prediction

This work presents an artificial intelligence (AI) framework for real-time, personalized sepsis prediction four hours before onset through fusion of electrocardiogram (ECG) and patient electronic medical record. An on-chip classifier combines analog reservoir-computer and artificial neural network to perform in-sensor classification at 43.6 TOPS/W (normalized efficiency of 528 TOPS/W) which reduces energy by 155× compared to conventional sensors and 4× compared to state-of-the-art bio-medical AI circuits. The proposed AI framework predicts sepsis onset with state-of-the-art 92.9% accuracy on patient data from MIMIC–III. The proposed framework is noninvasive and does not require lab tests which makes it suitable for at-home monitoring.