Research map / Architectures and efficiency / Pruning and efficient tuning
EmBrace: A Collective Knowledge Fusion Framework Toward Unified EEG Foundation Models
Ziyu Jia, Junyi Lin, Pu Wan, Jinxin Pi, Jingying Ma, Peiliang Gong, Xinliang Zhou, Yi Ding, Chenyu Liu
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 multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | LaBraM carrier · ISRUC-S3 | Kappa: 0.7035 -> 0.7098 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | LaBraM carrier · BCIC-IV-2A | Kappa: 0.1271 -> 0.1285 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | LaBraM carrier · MentalArithmetic | AUROC: 0.7073 -> 0.7234 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | CBraMod carrier · ISRUC-S3 | Kappa: 0.4463 -> 0.4678 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | CBraMod carrier · BCIC-IV-2A | Kappa: 0.3229 -> 0.3396 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | CBraMod carrier · MentalArithmetic | AUROC: 0.7979 -> 0.8060 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | CodeBrain carrier · ISRUC-S3 | Kappa: 0.5807 -> 0.6090 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | CodeBrain carrier · BCIC-IV-2A | Kappa: 0.1662 -> 0.1808 | Improved |
| add multi-EFM knowledge fusion on top of standalone fine-tuning (ablation) | CodeBrain carrier · MentalArithmetic | AUROC: 0.8487 -> 0.8813 | Improved |
| use EmBrace single carrier instead of deploying all three EFMs | EmBrace · FACED, ATTENTION, SEED-V, ISRUC-S1 | inference latency: 47.3%, 82.4%, 18.4%, 81.4% lower latency | Improved |
| use EmBrace instead of training all three EFMs separately | EmBrace · FACED, ATTENTION, SEED-V, ISRUC-S1 | training time per epoch: 4.1%, 46.7%, 0.6%, 57.4% lower training cost | Improved |
| EmBrace fusion versus best individual EFM on binary tasks | EmBrace · CHB-MIT | ACC-B: 0.7166 -> 0.6157 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- CBraMod — CBraMod: A Criss-Cross Brain Foundation Model for EEG Decoding (ICLR 2025) · better on 7 of 7 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- What do EEG Foundation Models Capture from Human Brain Signals? (NeurIPS 2026 · similarity 0.89)
- EEG Benchmarking Needs a Task Specification Layer: NeuroDoc for Rulebook-Guided, Executable Benchmark Construction (NeurIPS 2026 · similarity 0.88)
- Standing on the Shoulders of Giants: Rethinking EEG Foundation Model Pretraining via Multi-Teacher Distillation (NeurIPS 2026 · similarity 0.88)
- NeuroLM: A Universal Multi-task Foundation Model for Bridging the Gap between Language and EEG Signals (ICLR 2025 · similarity 0.87)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.87)
- Brain-OF: An Omnifunctional Foundation Model for fMRI, EEG and MEG (NeurIPS 2026 · 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.
