$Tiles$: Learning Multiple Time-Frequency Latents for Diverse Accelerometer Tasks

Wearable accelerometers support many applications, but the signal characteristics relevant to each task differ greatly. Existing self-supervised learning methods often encourage invariance to selected signal characteristics or aggregate each window into a single embedding, potentially discarding information required by some tasks. We present $Tiles$, a self-supervised method for learning transferable representations of accelerometer signals. $Tiles$ decomposes raw signals into frequency bands and uses iterative competitive cross-attention with learned queries to organize temporal and spectral features into multiple time-frequency latents. Two auxiliary reconstruction losses further encourage the latents to preserve local structure and high-frequency variations. To apply the learned representation to a target task, we keep the encoder fixed and train a lightweight prediction head that combines the latents. We evaluate $Tiles$ on 11 classification and regression tasks that broaden downstream accelerometer SSL evaluation beyond standard human activity recognition. $Tiles$ achieves the best result among the compared self-supervised methods on 8 tasks and performs comparably on the rest. Analysis of task-head weights and performance under individual-latent masking further shows that tasks rely on different combinations of the learned latents.