Research map / Vision-language and multimodal / Cross-modal learning
EEG-Based Multimodal Learning via Hyperbolic Mixture-of-Curvature Experts
Runhe Zhou, Shanglin Li, Guanxiang Huang, Xinliang Zhou, Qibin Zhao, Motoaki Kawanabe, Yi Ding, Cuntai Guan
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 |
|---|---|---|---|
| use hyperbolic encoder instead of Euclidean encoder (ablation) | EEG-MoCE · EAV | balanced accuracy (%): 60.33 -> 74.17 | Improved |
| add hyperbolic fusion on top of hyperbolic encoder (ablation) | EEG-MoCE · EAV | balanced accuracy (%): 74.17 -> 75.88 | Improved |
| add hyperbolic fusion with Euclidean encoder (ablation) | Euclidean baseline · EAV | balanced accuracy (%): 60.33 -> 61.48 | Improved |
| use learnable curvatures instead of fixed curvature K=-2 (ablation) | EEG-MoCE · EAV | balanced accuracy (%): +2.14% accuracy | Improved |
| add curvature-oriented multimodal fusion (curvature prior and scaled temperature) to attention (ablation) | EEG-MoCE · EAV | balanced accuracy (%): +1.38% accuracy | Improved |
| EEG-MoCE vs best baseline HEEGNet | EEG-MoCE · EAV | balanced accuracy (%): 61.74 -> 75.88 | Improved |
| EEG-MoCE vs XSleepFusion | EEG-MoCE · ISRUC-S3 | balanced accuracy (%): 75.19 -> 78.53 | Improved |
| EEG-MoCE vs EF-Net | EEG-MoCE · Cognitive N-back | balanced accuracy (%): 54.41 -> 62.39 | Improved |
| use hyperbolic operations instead of Euclidean fusion baselines (compute cost) | EEG-MoCE · EAV | train sec/epoch: 8.56 -> 26.07 | Hurt |
| encode raw modality with Euclidean encoder (hyperbolicity of features) | modality features · EAV EEG | delta_rel (lower is more tree-like): 0.097 -> 0.160 | Hurt |
Closest papers by meaning
- LAtte: Hyperbolic Lorentz Attention for Joint-Subject EEG Classification (NeurIPS 2026 · similarity 0.87)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.86)
- NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces (NeurIPS 2026 · similarity 0.86)
- What do EEG Foundation Models Capture from Human Brain Signals? (NeurIPS 2026 · similarity 0.86)
- B[FM]$^2$: Brain Foundation Model via Flow Matching with SplitUNet (NeurIPS 2026 · similarity 0.86)
- Geodesic Optimization for Predictive Shift Adaptation on EEG data (NeurIPS 2024 · similarity 0.86)
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.
