Sonata: A Hybrid World Model for Inertial Kinematics under Clinical Data Scarcity
For clinical researchers needing reliable kinematic models from small patient cohorts, Sonata provides a practical solution that achieves strong performance despite data scarcity.
Sonata introduces a compact latent world model for trunk IMU representation learning that outperforms a matched autoregressive baseline (MAE) in clinical discrimination, fall-risk prediction, and cross-cohort transfer across 14 evaluation tasks, using only 3.77M parameters for on-device inference.
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.