Separating Expert Retention from Autonomous Source Inference in Raw-ECG-Replay-Free Continual ECG Deployment
For researchers in continual learning for medical time-series, this work provides a clear decomposition of the problem into expert retention and source inference, showing that the latter is the key remaining challenge.
The paper studies continual ECG deployment where raw ECGs cannot be replayed, using frozen pretrained features and per-source classifiers. It finds that source-aware expert selection nearly matches an offline oracle (Macro-F1 0.7915 vs 0.7885), but autonomous source inference via a router leaves a gap of 0.011-0.0133, identifying this as the main bottleneck.
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 \ours{}, 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 $0.7915\pm0.0036$ Macro-F1 and a matched offline independent-head reference reaches $0.7885\pm0.0009$, supporting strong source-aware expert retention. Without source IDs, an MLP router reaches $0.7756\pm0.0027$ and top-2 margin fusion reaches $0.7782\pm0.0022$. The top-2 gain over hard MLP routing is small ($+0.0026$), with a 95\% confidence interval from paired bootstrap that includes zero. Across three domain orders, the top-2-to-oracle gap remains $0.0111$--$0.0133$, 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.