Research map / Learning theory and algorithms / Matrix and tensor methods
A Gyrovector Space Approach for Symmetric Positive Semi-Definite Matrix Learning
Xuan Son Nguyen
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 full-rank instead of low-rank learning (p=14) (ablation) | SPSD-AI · HDM05 | accuracy (%): 74.19 -> 81.32 | Improved |
| use full-rank instead of low-rank learning (p=14) (ablation) | SPSD-AI · NTU60 interaction (X-Sub) | accuracy (%): 89.78 -> 95.86 | Improved |
| use full-rank instead of low-rank learning (ablation) | SPSD-AI · HDM05 | training time per epoch (min): 0.66 -> 1.09 | Hurt |
| use full-rank instead of low-rank learning (ablation) | SPSD-AI · NTU60 interaction (X-Sub) | training time per epoch (min): 5.61 -> 9.19 | Hurt |
| add canonical representation in low-rank SPSD layer (ablation) | SPSD-AI low-rank · HDM05 | accuracy (%): 41.37 -> 74.19 | Improved |
| add canonical representation in low-rank SPSD layer (ablation) | SPSD-AI low-rank · FPHA | accuracy (%): 63.41 -> 90.94 | Improved |
| add canonical representation in low-rank SPSD layer (ablation) | SPSD-AI low-rank · NTU60 interaction (X-Sub) | accuracy (%): 56.82 -> 89.78 | Improved |
| add canonical representation in low-rank SPSD layer (ablation) | SPSD-AI low-rank · NTU60 interaction (X-View) | accuracy (%): 58.34 -> 90.15 | Improved |
| use Affine-invariant geometry instead of Log-Euclidean | SPSD networks · HDM05 | accuracy (%): 77.46 -> 81.32 | Improved |
| use Affine-invariant geometry instead of Log-Cholesky | SPSD networks · HDM05 | accuracy (%): 73.52 -> 81.32 | Improved |
| use Affine-invariant geometry instead of Log-Euclidean | SPSD networks · NTU60 interaction (X-Sub) | accuracy (%): 90.74 -> 95.86 | Improved |
| use Affine-invariant geometry instead of Log-Cholesky | SPSD networks · FPHA | accuracy (%): 89.73 -> 96.58 | Improved |
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.
- GeomNet — GeomNet: A Neural Network Based on Riemannian Geometries of SPD Matrix Space and Cholesky Space for 3D Skeleton-Based Interaction Recognition (ICCV 2021) · better on 2 of 2 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
- Matrix Manifold Neural Networks++ (ICLR 2024 · similarity 0.91)
- Neural networks on Symmetric Spaces of Noncompact Type (ICLR 2025 · similarity 0.87)
- Riemannian Multinomial Logistics Regression for SPD Neural Networks (CVPR 2024 · similarity 0.87)
- Siegel Neural Networks (NeurIPS 2025 · similarity 0.87)
- Understanding Matrix Function Normalizations in Covariance Pooling through the Lens of Riemannian Geometry (ICLR 2025 · similarity 0.86)
- RMLR: Extending Multinomial Logistic Regression into General Geometries (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.
