Research map / Architectures and efficiency / Attention and transformers
BIOT: Biosignal Transformer for Cross-data Learning in the Wild
Chaoqi Yang, M Brandon Westover, Jimeng Sun
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 unsupervised pre-training on PREST (ablation) | BIOT · CHB-MIT | Balanced Acc.: 0.6640 -> 0.6942 | Improved |
| add unsupervised pre-training on PREST (ablation) | BIOT · TUEV | Balanced Acc.: 0.4682 -> 0.5207 | Improved |
| add unsupervised pre-training on PREST (ablation) | BIOT · TUAB | Balanced Acc.: 0.7925 -> 0.7907 | No clear effect |
| add PREST+SHHS joint pre-training (ablation) | BIOT · TUAB | Balanced Acc.: 0.7925 -> 0.8019 | Improved |
| add PREST+SHHS joint pre-training (ablation) | BIOT · IIIC Seizure | Balanced Acc.: 0.5762 -> 0.5800 | Improved |
| add ECG pre-training on Cardiology with 12 leads (ablation) | BIOT · PTB-XL | Balanced Acc.: 0.8315 -> 0.8421 | Improved |
| pretrain with 12 leads instead of 6 leads (ablation) | BIOT Cardiology · PTB-XL | Balanced Acc.: 0.8350 -> 0.8421 | Improved |
| supervised pre-training on IIIC Seizure (16ch, 10s) (ablation) | BIOT · TUEV | Cohen's Kappa: 0.4482 -> 0.5316 | Improved |
| supervised pre-training on CHB-MIT (8ch, 10s) (ablation) | BIOT · TUEV | Balanced Acc.: 0.4682 -> 0.4123 | Hurt |
| use 16 channels and 10s instead of 5s or 8 channels in supervised pretraining (ablation) | BIOT on TUAB pretrain · TUEV | Balanced Acc.: 0.4954 -> 0.5256 | Improved |
| pre-train on all six EEG datasets vs PREST+SHHS (ablation) | BIOT · IIIC Seizure | Weighted F1: 0.5878 -> 0.5737 | Mixed |
| pre-training with linear-attention BIOT vs scratch under missing channels/segments (ablation) | BIOT · TUEV, IIIC Seizure | Kappa/Weighted F1 | Improved |
Closest papers by meaning
- CiTrus: Squeezing Extra Performance out of Low-data Bio-signal Transfer Learning (AAAI 2025 · similarity 0.88)
- EEGPT: Pretrained Transformer for Universal and Reliable Representation of EEG Signals (NeurIPS 2024 · similarity 0.87)
- NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces (NeurIPS 2026 · similarity 0.85)
- MindAlign: Bridging EEG, Vision, and Language for Zero-Shot Visual Decoding (NeurIPS 2026 · similarity 0.84)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.84)
- Medformer: A Multi-Granularity Patching Transformer for Medical Time-Series Classification (NeurIPS 2024 · similarity 0.84)
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.
