Research map / Optimization / Implicit bias and generalization
A Riemannian Network for SPD Matrix Learning
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 |
|---|---|---|---|
| stack additional BiMap/ReEig blocks (ablation) | SPDNet · AFEW validation | accuracy: 26.32% -> 34.23% | Helped |
| stack additional BiMap/ReEig blocks (ablation) | SPDNet · HDM05 | accuracy: 48.12%$\pm$3.15 -> 61.45%$\pm$1.12 | Helped |
| stack additional BiMap/ReEig blocks (ablation) | SPDNet · PaSC1 control | accuracy: 68.52% -> 80.12% | Helped |
| stack additional BiMap/ReEig blocks (ablation) | SPDNet · PaSC2 handheld | accuracy: 63.92% -> 72.83% | Helped |
| remove the LogEig layer (ablation) | SPDNet · AFEW validation | accuracy: 34.23% -> 21.49% | Hurt |
| remove the LogEig layer (ablation) | SPDNet · HDM05 | accuracy: 61.45%$\pm$1.12 -> 4.89% | Hurt |
| apply the ReEig activation at ε=10^{-4} rather than ε=0 (ablation) | SPDNet-3BiRe · AFEW validation | accuracy: 32.35% -> 34.23% | Helped |
| apply the ReEig activation at ε=5 × 10^{-5} rather than ε=0 (ablation) | SPDNet-3BiRe · AFEW validation | accuracy: 32.35% -> 33.15% | Helped |
Closest papers by meaning
- Riemannian Multinomial Logistics Regression for SPD Neural Networks (CVPR 2024 · similarity 0.89)
- Building Transformation Layers for Riemannian Neural Networks (NeurIPS 2026 · similarity 0.89)
- Learning to Normalize on the SPD Manifold under Bures-Wasserstein Geometry (CVPR 2025 · similarity 0.89)
- Matrix Manifold Neural Networks++ (ICLR 2024 · similarity 0.88)
- RMLR: Extending Multinomial Logistic Regression into General Geometries (NeurIPS 2024 · similarity 0.88)
- Understanding Matrix Function Normalizations in Covariance Pooling through the Lens of Riemannian Geometry (ICLR 2025 · 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.