Evidence-Guided Neural Architecture Selection under Uncertainty for Subject-Specific Blood Glucose Forecasting

2026-06-03

AI Summary

This work addresses the challenge of personalized blood glucose prediction under conditions of scarce, noisy, and highly heterogeneous data by proposing the EVIDENT framework. EVIDENT uniquely integrates Dempster–Shafer evidence theory with Bayesian uncertainty modeling into neural architecture selection. The method employs evidence-driven ranking and task-oriented validation to automatically identify the smallest-capacity model from a candidate pool that meets performance criteria, while enabling rationale-weighted ensemble predictions. Experimental results demonstrate that, compared to random search baselines, models selected by EVIDENT are significantly more lightweight and yield more stable predictions, achieving markedly improved generalization on unseen patients and effectively mitigating the risks of both under- and over-parameterization.

Abstract

Reliable neural architecture selection is an open challenge in time-series forecasting under limited, noisy, and heterogeneous data, where standard heuristic architecture design and validation approaches fail to ensure accurate and reliable prediction and generalization. We propose EVIDENT (EVidence-based IDEntification of Neural archiTectures), a framework for architecture selection that integrates Bayesian training, evidence-based ranking, and task-specific validation under uncertainty. The framework explores the candidate architecture pool and identifies the lowest-capacity model that satisfies a prescribed validation criterion. We demonstrate this method using temporal convolutional networks (TCNs) for individualized blood glucose forecasting in type 1 diabetes patients. The results show that EVIDENT systematically rejects both under- and over-parameterized TCN architectures on population-level diabetes data, while identifying models that generalize reliably to unseen patients. When multiple architectures are competitive, the framework further supports plausibility-weighted ensemble predictions that enhance predictive performance. Compared with a random-search baseline, EVIDENT identified smaller architectures with more consistent forecasting performance on unseen patients. These findings establish EVIDENT as a strategy to neural architecture discovery, enabling reliable model selection for high-consequence forecasting in data-limited and heterogeneous settings.

Problem

• neural architecture selection
• time-series forecasting
• blood glucose prediction
• data heterogeneity
• model generalization

Innovation

• neural architecture selection
• Bayesian training
• evidence-based ranking
• temporal convolutional networks
• uncertainty-aware validation