Sonata: A Hybrid World Model for Inertial Kinematics under Clinical Data Scarcity

TL;DR

Sonata is a compact hybrid world model trained on clinical data that improves inertial kinematic predictions and clinical assessments with limited data, suitable for on-device wearable use.

Contribution

A novel 3.77M-parameter hybrid model pre-trained on multiple datasets, optimized for clinical inertial data scarcity and improved predictive performance.

Findings

Outperforms autoregressive baselines in clinical discrimination tasks.

Yields more structured and higher-rank latent representations.

Enables on-device wearable inference for neurological assessment.

Abstract

We introduce Sonata, a compact latent world model for six-axis trunk IMU representation learning under clinical data scarcity. Clinical cohorts typically comprise tens to hundreds of patients, making web-scale masked-reconstruction objectives poorly matched to the problem. Sonata is a 3.77 M-parameter hybrid model, pre-trained on a harmonised corpus of nine public datasets (739 subjects, 190k windows) with a latent world-model objective that predicts future state rather than reconstructing raw sensor traces. In a controlled comparison against a matched autoregressive forecasting baseline (MAE) on the same backbone, Sonata yields consistently stronger frozen-probe clinical discrimination, prospective fall-risk prediction, and cross-cohort transfer across a 14-arm evaluation suite, while producing higher-rank, more structured latent representations. At 3.77 M parameters the model is compatible with on-device wearable inference, offering a step toward general kinematic world models for neurological assessment.