The Driver-Blindness Phenomenon: Why Deep Sequence Models Default to Autocorrelation in Blood Glucose Forecasting
This addresses a critical problem for diabetes management by exposing why models default to autocorrelation, potentially improving forecasting accuracy and patient outcomes, though the solutions are incremental.
The paper identifies Driver-Blindness, where deep sequence models for blood glucose forecasting fail to use clinically informative drivers like insulin and meals, showing near-zero performance gain (Δ_drivers) over univariate baselines. It attributes this to architectural biases, data fidelity gaps, and physiological heterogeneity, and proposes mitigation strategies such as feature encoders and personalization.
Deep sequence models for blood glucose forecasting consistently fail to leverage clinically informative drivers--insulin, meals, and activity--despite well-understood physiological mechanisms. We term this Driver-Blindness and formalize it via $Δ_{\text{drivers}}$, the performance gain of multivariate models over matched univariate baselines. Across the literature, $Δ_{\text{drivers}}$ is typically near zero. We attribute this to three interacting factors: architectural biases favoring autocorrelation (C1), data fidelity gaps that render drivers noisy and confounded (C2), and physiological heterogeneity that undermines population-level models (C3). We synthesize strategies that partially mitigate Driver-Blindness--including physiological feature encoders, causal regularization, and personalization--and recommend that future work routinely report $Δ_{\text{drivers}}$ to prevent driver-blind models from being considered state-of-the-art.