Research map / Probabilistic and scientific modeling / Dynamical systems
SimGANs: Simulator-Based Generative Adversarial Networks for ECG Synthesis to Improve Deep ECG Classification
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 simulator Euler loss to DCGAN-generated training data | ResNet heartbeat classifier · MIT-BIH SVEB (S) | precision: 0.58 -> 0.80 | Helped |
| Add simulator Euler loss to DCGAN-generated training data | ResNet heartbeat classifier · MIT-BIH VEB (V) | precision: 0.72 -> 0.84 | Helped |
| Add simulator Euler loss to DCGAN-generated training data | ResNet heartbeat classifier · MIT-BIH Fusion (F) | precision: 0.20 -> 0.40 | Helped |
| Add simulator Euler loss to vanilla-GAN-generated training data | ResNet heartbeat classifier · MIT-BIH SVEB (S) | precision: 0.63 -> 0.64 | Helped |
| Add simulator Euler loss to vanilla-GAN-generated training data | ResNet heartbeat classifier · MIT-BIH VEB (V) | precision: 0.70 -> 0.82 | Helped |
| Add simulator Euler loss to vanilla-GAN-generated training data | ResNet heartbeat classifier · MIT-BIH Fusion (F) | precision: 0.10 -> 0.25 | Helped |
| Add synthetic ECG training data from any generative model | ResNet heartbeat classifier · MIT-BIH arrhythmia classes | classification performance | Helped |
| Use simulator-generated ECG as generator input and omit Euler loss | DCGAN-based RefineGAN · MIT-BIH SVEB (S) | precision: 0.47 -> 0.7 | Hurt |
| Use simulator-generated ECG as generator input and omit Euler loss | DCGAN-based RefineGAN · MIT-BIH VEB (V) | precision: 0.73 -> 0.89 | Hurt |
| Add synthetic ECG directly from the simulator instead of learned SimDCGAN samples (ablation) | ResNet heartbeat classifier · MIT-BIH SVEB and Fusion | precision-recall performance | Hurt |
| Remove simulator-guided generation and use an ordinary DCGAN (ablation) | ResNet heartbeat classifier · MIT-BIH SVEB (S) | precision: 0.80 vs. 0.58 | Hurt |
| Remove simulator-guided generation and use an ordinary DCGAN (ablation) | ResNet heartbeat classifier · MIT-BIH VEB (V) | precision: 0.84 vs. 0.72 | Hurt |
Closest papers by meaning
- Leveraging an ECG Beat Diffusion Model for Morphological Reconstruction from Indirect Signals (NeurIPS 2024 · similarity 0.85)
- Robustness and Visual Explanation for Black Box Image, Video, and ECG Signal Classification with Reinforcement Learning (AAAI 2024 · similarity 0.85)
- Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model (ICLR 2025 · similarity 0.83)
- Boosting Masked ECG-Text Auto-Encoders as Discriminative Learners (ICML 2025 · similarity 0.82)
- ECHOPulse: ECG Controlled Echocardio-gram Video Generation (ICLR 2025 · similarity 0.82)
- From Token to Rhythm: A Multi-Scale Approach for ECG-Language Pretraining (ICML 2025 · similarity 0.82)
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.