Research map / Architectures and efficiency / RNNs and normalization
Riemannian batch normalization for SPD neural networks
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 Riemannian batch normalization after each BiMap layer | SPDNet · NATO radar, all training data | accuracy: 85.4% -> 87.2% | Helped |
| add Riemannian batch normalization | SPDNet · NATO radar, 10% training data | accuracy: 83.9% -> 84.7% | Helped |
| add Riemannian batch normalization | SPDNet · AFEW | accuracy: 29.9 -> 34.9 ({400,50}); 33.7 -> 37.1 (deepest) | Helped |
| add Riemannian batch normalization | SPDNet · HDM05 {93,30} | accuracy: 61.6% -> 68.1% | Helped |
| add Riemannian batch normalization | SPDNet deepest net · AFEW | training time (s): 81s -> 88s (+8.6%) | Hurt |
| use SPDNetBN (~500 params) instead of full-size FCN (~10000 params) | NATO radar · NATO radar, all training data | accuracy: 89.4% -> 87.2% | Hurt |
| use SPDNetBN instead of full-size FCN with 10% training data | NATO radar · NATO radar, 10% training data | accuracy: 66.6% -> 84.7% | Helped |
Closest papers by meaning
- A Lie Group Approach to Riemannian Batch Normalization (ICLR 2024 · similarity 0.91)
- Understanding Matrix Function Normalizations in Covariance Pooling through the Lens of Riemannian Geometry (ICLR 2025 · similarity 0.90)
- Learning to Normalize on the SPD Manifold under Bures-Wasserstein Geometry (CVPR 2025 · similarity 0.90)
- Riemannian Multinomial Logistics Regression for SPD Neural Networks (CVPR 2024 · similarity 0.88)
- RMLR: Extending Multinomial Logistic Regression into General Geometries (NeurIPS 2024 · similarity 0.88)
- Matrix Manifold Neural Networks++ (ICLR 2024 · similarity 0.88)
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.