Research map / Label-efficient and robust learning / Recognition and long-tail
VTaC: A Benchmark Dataset of Ventricular Tachycardia Alarms from ICU Monitors
Li-wei H. Lehman, Benjamin E Moody, Harsh Deep, Feng Wu, Hasan Saeed, Lucas McCullum, Diane Perry, Tristan Struja, Qiao Li, Gari Clifford, Roger Mark
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 contrastive learning to 1-D CNN | CNN · VTaC real-time | Challenge Score: 76.17 -> 79.07 | Improved |
| add contrastive learning to FCN | FCN · VTaC real-time | Challenge Score: 80.08 -> 78.41 | Hurt |
| add contrastive learning to FCN | FCN · VTaC retrospective | Challenge Score: 75.84 -> 78.95 | Improved |
| add contrastive learning to CNN | CNN · VTaC retrospective | Challenge Score: 76.94 -> 78.85 | Improved |
| add 5 seconds of post-alarm data (retrospective vs real-time) | CNN · VTaC | Challenge Score: 76.17 -> 76.94 | No clear effect |
| add 5 seconds of post-alarm data (retrospective vs real-time) | FCN+CL · VTaC | Challenge Score: 78.41 -> 78.95 | Improved |
| add 5 seconds of post-alarm data (retrospective vs real-time) | FCN · VTaC | Challenge Score: 80.08 -> 75.84 | Hurt |
| replace rule-based method by FCN | Rule-based (Plesinger) · VTaC real-time | Challenge Score: 67.32 -> 80.08 | Improved |
| replace CNN by FCN as feature extractor | CNN · VTaC real-time | AUC: 0.936 -> 0.949 | Improved |
| replace supervised CNN by Transformer encoder | CNN · VTaC real-time | Challenge Score: 76.17 -> 62.73 | Hurt |
| use generative anomaly detection (Diffusion+CL) instead of supervised FCN | FCN · VTaC real-time | Challenge Score: 80.08 -> 52.51 | Hurt |
| use autoencoder reconstruction loss (SAE) instead of plain MLP | MLP · VTaC real-time | Challenge Score: 45.58 -> 68.77 | Improved |
Closest papers by meaning
- Yet Another ICU Benchmark: A Flexible Multi-Center Framework for Clinical ML (ICLR 2024 · similarity 0.82)
- Reading Your Heart: Learning ECG Words and Sentences via Pre-training ECG Language Model (ICLR 2025 · similarity 0.81)
- GEM: Empowering MLLM for Grounded ECG Understanding with Time Series and Images (NeurIPS 2025 · similarity 0.81)
- Parallel-Learning of Invariant and Tempo-variant Attributes of Single-Lead Cardiac Signals: PLITA (AAAI 2025 · similarity 0.81)
- From Token to Rhythm: A Multi-Scale Approach for ECG-Language Pretraining (ICML 2025 · similarity 0.81)
- ECG-Reasoning-Benchmark: A Benchmark for Evaluating Clinical Reasoning Capabilities in ECG Interpretation (NeurIPS 2026 · similarity 0.80)
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.
