HiMAE: Hierarchical Masked Autoencoders Discover Resolution-Specific Structure in Wearable Time Series

ICLR 2026 Conference Submission20543 Authors

Keywords: SSL, Wearables, Interpretability, Inductive Bias

TL;DR: We propose a lightweight SSL objective that competes with much larger transformer Foundation models that also serve as an interpretability tool.

Abstract: Wearable sensors provide abundant physiological time series, yet the principles governing their predictive utility remain unclear. We hypothesize that temporal resolution is a fundamental axis of representation learning, with different clinical and behavioral outcomes relying on structure at distinct scales. To test this \emph{resolution hypothesis}, we introduce \emph{HiMAE} (Hierarchical Masked Autoencoder), a self-supervised framework that combines masked autoencoding with a hierarchical convolutional encoder–decoder. HiMAE produces multi-resolution embeddings that enable systematic evaluation of which temporal scales carry predictive signal, transforming resolution from a hyperparameter into a probe for interpretability. Across classification, regression, and generative benchmarks, HiMAE consistently outperforms state-of-the-art foundation models that collapse scale, while being orders of magnitude smaller. HiMAE is an efficient representation learner compact enough to run entirely on-watch, achieving sub-millisecond inference on smartwatch-class CPUs for true edge inference. Together, these contributions position HiMAE as both an efficient self supervised learning method and a discovery tool for scale-sensitive structure in wearable health.

Primary Area: applications to physical sciences (physics, chemistry, biology, etc.)

Submission Number: 20543