Separating Expert Retention from Autonomous Source Inference in Raw-ECG-Replay-Free Continual ECG Deployment
arXiv:2607.01674
The paper proposes a replay‑free continual learning framework for multi‑source ECG classification that keeps a frozen pretrained backbone, adds a new linear expert for each incoming domain, and uses a lightweight router with top‑2 margin fusion to infer the source when metadata are unavailable.
Abstract
In multi-source ECG deployment, models may need to incorporate new data sources when earlier raw ECGs cannot be retained or replayed. Freezing a pretrained backbone and assigning each source an isolated classifier prevents parameter interference, but deployment still requires selecting an expert when source metadata are unavailable. We study this distinction through IRFE-ECG, an incremental expert bank built on frozen 1024-dimensional ECGFounder features. Each arriving domain adds a balanced-softmax linear expert, while a lightweight router is fitted only on retained training features and domain labels from sources observed so far. A validation-calibrated margin rule fuses the two most likely experts instead of committing to a single routed expert. On CPSC, PTB-XL, Georgia, and Chapman-Shaoxing, source-aware expert selection reaches Macro-F1 and a matched offline independent-head reference reaches , supporting strong source-aware expert retention. Without source IDs, an MLP router reaches and top-2 margin fusion reaches . The top-2 gain over hard MLP routing is small (), with a 95\% confidence interval from paired bootstrap that includes zero. Across three domain orders, the top-2-to-oracle gap remains --, identifying autonomous source inference as the main remaining bottleneck. No raw ECGs are replayed, but frozen training features are retained for router updates; the method is therefore not memory-free.Code is available at https://github.com/yufanlu221/IRFE-ECG.
Submitted to BIBM 2026