Research map / Label-efficient and robust learning / Recognition and long-tail
Similarity Learning via Kernel Preserving Embedding
Zhao Kang, Yiwei Lu, Yuanzhang Su, Changsheng Li, Zenglin Xu
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 |
|---|---|---|---|
| reconstruct kernel matrix instead of raw data (SLKE-R vs LRR) | low-rank self-expression · 9 clustering datasets | accuracy: 8.92% average accuracy gain | Improved |
| reconstruct kernel matrix instead of raw data (SLKE-S vs SSC) | sparse self-expression · 9 clustering datasets | accuracy: 8.76% average accuracy gain | Improved |
| kernel-preserving similarity vs raw-data LRR | LRR · YALE | accuracy: 61.21 -> 66.24 | Improved |
| kernel-preserving similarity vs raw-data SSC | SSC · TDT2 | accuracy: 23.45 -> 74.92 | Improved |
| SLKE vs adaptive-neighbor CAN | SLKE-R · COIL20 | accuracy: 84.58 -> 84.03 | Hurt |
| SLKE vs adaptive-neighbor CAN | SLKE-R · TR41 | accuracy: 62.87 -> 74.37 | Improved |
| SLKE vs TLSC | SLKE-R · TDT2 | accuracy: 55.74 -> 74.92 | Improved |
| low-rank (SLKE-R) vs sparse (SLKE-S) regularizer | SLKE · ORL / BA | accuracy: ORL 77.00 vs 74.75; BA 50.74 vs 44.37 | Mixed |
| best kernel vs average over 12 kernels | SLKE-R · YALE | accuracy: 51.28 -> 66.24 | Improved |
| vary gamma from 1e-6 to 1e-3 (ablation) | SLKE · YALE and JAFFE, Gaussian kernel t=100 | accuracy / NMI | No clear effect |
| Wilcoxon test of SLKE-R vs SC | SLKE-R · all 9 datasets | accuracy: p = .0039 | Improved |
| Wilcoxon test of SLKE-R vs LRR | SLKE-R · all 9 datasets | accuracy: p = .0977 | No clear effect |
Closest papers by meaning
- Twin Learning for Similarity and Clustering: A Unified Kernel Approach (AAAI 2017 · similarity 0.89)
- Fast Approximation of Similarity Graphs with Kernel Density Estimation (NeurIPS 2023 · similarity 0.87)
- Towards an Explainable Comparison and Alignment of Feature Embeddings (ICML 2025 · similarity 0.86)
- Stabilizing Sample Similarity in Representation via Mitigating Random Consistency (ICML 2025 · similarity 0.85)
- Learning to Embed Distributions via Maximum Kernel Entropy (NeurIPS 2024 · similarity 0.85)
- Dynamic Similarity Graph Construction with Kernel Density Estimation (ICML 2025 · similarity 0.84)
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.
