Beyond Patient Invariance: Learning Cardiac Dynamics via Action-Conditioned JEPAs
Jose Geraldo Fernandes, Luiz Facury, Pedro Robles Dutenhefner, Wagner Meira
TLDR
A new self-supervised learning approach, Action-Conditioned JEPAs, models cardiac disease progression dynamically, improving triage and sample efficiency.
Key contributions
- Proposes Action-Conditioned World Models to simulate disease progression in physiological time-series, moving beyond static invariance.
- Adapts LeJEPA to define pathology as a transition vector acting on a patient's latent state, disentangling features.
- Predicts future electrophysiological states given disease onset, separating stable anatomy from dynamic pathological forces.
- Outperforms supervised baselines on cardiac triage and shows superior sample efficiency on the MIMIC-IV-ECG dataset.
Why it matters
This work shifts self-supervised learning in healthcare from static invariance to dynamic disease modeling. By simulating progression, it better aligns with clinical diagnosis, enabling more robust and sample-efficient detection of transient pathologies. This could lead to improved early detection and triage in low-resource settings.
Original Abstract
Self-supervised learning in healthcare has largely relied on invariance-based objectives, which maximize similarity between different views of the same patient. While effective for static anatomy, this paradigm is fundamentally misaligned with clinical diagnosis, as it mathematically compels the model to suppress the transient pathological changes it is intended to detect. We propose a shift towards Action-Conditioned World Models that learn to simulate the dynamics of disease progression, or Event-Conditioned. Adapting the LeJEPA framework to physiological time-series, we define pathology not as a static label, but as a transition vector acting on a patient's latent state. By predicting the future electrophysiological state of the heart given a disease onset, our model explicitly disentangles stable anatomical features from dynamic pathological forces. Evaluated on the MIMIC-IV-ECG dataset, our approach outperforms fully supervised baselines on the critical triage task. Crucially, we demonstrate superior sample efficiency: in low-resource regimes, our world model outperforms supervised learning by over 0.05 AUROC. These results suggest that modeling biological dynamics provides a dense supervision signal that is far more robust than static classification. Source code is available at https://github.com/cljosegfer/lesaude-dynamics
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.