Research map / Label-efficient and robust learning / Test-time adaptation
Convolution Monge Mapping Normalization for learning on sleep data
Theo Gnassounou, Rémi Flamary, Alexandre Gramfort
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 CMMN instead of Sample normalization | Chambon · Cross-dataset mean | BACC (%): 58.6 ± 2.6 -> 62.7 ± 2.4 | Improved |
| Use CMMN instead of Sample normalization | Chambon · SHHS→MASS | BACC (%): 57.0 ± 2.8 -> 64.5 ± 2.8 | Improved |
| Use CMMN instead of no normalization | Chambon · Phys.→SHHS | BACC (%): 53.9 ± 3.2 -> 51.6 ± 1.8 | Hurt |
| Use CMMN vs subject-specific normalizations (Conv, Norm, ConvNorm) | Chambon · Physionet cross-session | BACC (%): 68.1 ± 1.3 -> 74.8 ± 0.6 | Improved |
| Use subject-specific normalization instead of No Adapt | Chambon · Physionet cross-session | BACC (%): 73.7 ± 0.7 -> 67.5 ± 2.7 | Hurt |
| Use CMMN with DeepSleepNet | DeepSleepNet · Mean (9 pairs) | BACC (%): 54.6 ± 2.9 -> 58.6 ± 3.2 | Improved |
| Use CMMN with DeepSleepNet | DeepSleepNet · MASS→MASS | BACC (%): 73.3 ± 1.7 -> 73.1 ± 2.6 | No clear effect |
| Add DANN on top of CMMN | CMMN · Cross-dataset mean (6 pairs) | BACC (%): 60.0 ± 2.7 -> 59.6 ± 1.7 | No clear effect |
| Add DANN on top of CMMN | CMMN · Cross-dataset mean (6 pairs) | ΔBACC@20 (%): 8.5 ± 4.4 -> 9.1 ± 4.9 | Improved |
| Add DANN alone to No Adapt | Chambon · Cross-dataset mean (6 pairs) | BACC (%): 54.9 ± 4.5 -> 55.4 ± 1.6 | No clear effect |
| Map to power-law PSD instead of barycenter (ablation) | CMMN · Mean (9 pairs), PSD target | BACC (%): 63.8 ± 2.9 -> 61.9 ± 6.4 | Hurt |
| Map to white-noise PSD (whitening) instead of barycenter (ablation) | CMMN · Mean (9 pairs), PSD target | BACC (%): 63.8 ± 2.9 -> 61.9 ± 5.1 | Hurt |
Closest papers by meaning
- SPDIM: Source-Free Unsupervised Conditional and Label Shift Adaptation in EEG (ICLR 2025 · similarity 0.85)
- SleepFM: Multi-modal Representation Learning for Sleep Across Brain Activity, ECG and Respiratory Signals (ICML 2024 · similarity 0.85)
- Generalizable Sleep Staging via Multi-Level Domain Alignment (AAAI 2024 · similarity 0.84)
- BrainOmni: A Brain Foundation Model for Unified EEG and MEG Signals (NeurIPS 2025 · similarity 0.84)
- NeuroAtlas: Benchmarking Foundation Models for Clinical EEG and Brain-Computer Interfaces (NeurIPS 2026 · similarity 0.84)
- Personalized Sleep Staging Leveraging Source-free Unsupervised Domain Adaptation (AAAI 2025 · similarity 0.83)
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.
