Research map / Image generation / Fast diffusion sampling
SE-Diff: Simulator and Experience Enhanced Diffusion Model for Comprehensive ECG Generation
Xiaoda Wang, Kaiqiao Han, Yuhao Xu, Xiao Luo, Yizhou Sun, Wei Wang, Carl Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add experience retrieval conditioning (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | MAE_HR: 15.06 -> 8.43 | Improved |
| add experience retrieval conditioning (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | rCLIP: 0.9099 -> 0.9470 | Improved |
| add experience retrieval conditioning (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | MAE: 0.0926 -> 0.0923 | No clear effect |
| add inter-lead constraint (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | MAE_HR: 19.21 -> 8.43 | Improved |
| add inter-lead constraint (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | rFID: 0.9128 -> 0.9509 | Improved |
| add simulator Euler loss (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | MAE: 0.0965 -> 0.0923 | Improved |
| add simulator Euler loss (ablation) | SE-Diff · MIMIC-IV-ECG text-to-ECG generation | NRMSE: 0.0768 -> 0.0714 | Improved |
| SE-Diff instead of DiffuSETS | text-to-ECG generator · MIMIC-IV-ECG text-to-ECG generation | MAE_HR: 13.29 -> 8.43 | Improved |
| SE-Diff instead of DiffuSETS | text-to-ECG generator · MIMIC-IV-ECG text-to-ECG generation | MAE: 0.1092 -> 0.0923 | Improved |
| augment minority class with SE-Diff ECGs instead of DiffuSETS | MLP classifier · Sex classification (Male = 10% Female) | F1 (%): 44 -> 58 | Improved |
| augment minority class with SE-Diff ECGs instead of DiffuSETS | MLP classifier · Sinus vs SVT (SVT = 10% Sinus) | F1 (%): 70 -> 72 | Improved |
| augment minority class with SE-Diff ECGs vs no augmentation | MLP classifier · Sinus vs SVT (SVT = 10% Sinus) | F1 (%): 56 -> 72 | Improved |
Closest papers by meaning
- Leveraging an ECG Beat Diffusion Model for Morphological Reconstruction from Indirect Signals (NeurIPS 2024 · similarity 0.90)
- Extending Pretrained 10-Second ECG Foundation Models to Longer Horizons (NeurIPS 2026 · similarity 0.88)
- Region-Disentangled Diffusion Model for High-Fidelity PPG-to-ECG Translation (AAAI 2024 · similarity 0.87)
- From Token to Rhythm: A Multi-Scale Approach for ECG-Language Pretraining (ICML 2025 · similarity 0.87)
- ECHOPulse: ECG Controlled Echocardio-gram Video Generation (ICLR 2025 · similarity 0.87)
- Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model (ICLR 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
