Pretraining ECG Data with Adversarial Masking Improves Model Generalizability for Data-Scarce Tasks

Open in new window