Uncertainty quantification with approximate variational learning for wearable photoplethysmography prediction tasks

2025-05-16

AI Summary

This work addresses the poor reliability, miscalibration, weak generalization, and low clinical trustworthiness of deep models when using wearable photoplethysmography (PPG) signals for atrial fibrillation (AF) classification and blood pressure regression. We conduct the first systematic comparison of Monte Carlo Dropout and an improved variational online Newton (IVON) method for uncertainty decomposition in PPG time-series modeling. We propose a calibration evaluation framework that decouples epistemic from aleatoric uncertainty, uncovering a dynamic hyperparameter-dependent regulation mechanism governing their relative proportions, and identify the necessity of class-stratified calibration assessment. Experiments demonstrate substantial improvements in uncertainty calibration quality. Our approach establishes a novel, interpretable, and verifiable paradigm for trustworthy AI in wearable health monitoring.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Probabilistic Programming
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Photoplethysmography (PPG) signals encode information about relative changes in blood volume that can be used to assess various aspects of cardiac health non-invasively, e.g. to detect atrial fibrillation (AF) or predict blood pressure (BP). Deep networks are well-equipped to handle the large quantities of data acquired from wearable measurement devices. However, they lack interpretability and are prone to overfitting, leaving considerable risk for poor performance on unseen data and misdiagnosis. Here, we describe the use of two scalable uncertainty quantification techniques: Monte Carlo Dropout and the recently proposed Improved Variational Online Newton. These techniques are used to assess the trustworthiness of models trained to perform AF classification and BP regression from raw PPG time series. We find that the choice of hyperparameters has a considerable effect on the predictive performance of the models and on the quality and composition of predicted uncertainties. E.g. the stochasticity of the model parameter sampling determines the proportion of the total uncertainty that is aleatoric, and has varying effects on predictive performance and calibration quality dependent on the chosen uncertainty quantification technique and the chosen expression of uncertainty. We find significant discrepancy in the quality of uncertainties over the predicted classes, emphasising the need for a thorough evaluation protocol that assesses local and adaptive calibration. This work suggests that the choice of hyperparameters must be carefully tuned to balance predictive performance and calibration quality, and that the optimal parameterisation may vary depending on the chosen expression of uncertainty.

Problem

• Quantify uncertainty in wearable PPG prediction tasks
• Improve interpretability and reduce overfitting in deep networks
• Assess trustworthiness of AF classification and BP regression models

Innovation

• Uses Monte Carlo Dropout for uncertainty quantification
• Applies Improved Variational Online Newton technique
• Optimizes hyperparameters for predictive performance calibration