Research map / Label-efficient and robust learning / Recognition and long-tail
Probabilistic Structural Latent Representation for Unsupervised Embedding
Mang Ye, Jianbing Shen
Check 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 |
|---|---|---|---|
| Add PSLR to AND neighborhood discovery | AND · CIFAR-10 | kNN accuracy (k=200, %): 86.3 -> 88.4 | Improved |
| Use PSLR rather than AND | Unsupervised embedding learning · STL-10 | kNN accuracy (%): 80.2 -> 83.2 | Improved |
| Use PSLR rather than AND | Unsupervised embedding learning · CUB200 | R@1 (%): 47.3 -> 48.1 | Improved |
| Use PSLR rather than ISIF | Unsupervised embedding learning · Car196 | R@1 (%): 41.3 -> 43.7 | Improved |
| Use PSLR rather than ISIF | Unsupervised embedding learning · Product | R@1 (%): 48.9 -> 51.1 | Improved |
| Add smooth variational self-reconstruction loss (ablation) | PSLR using Lz · Car196 | R@1 (%): 42.1 -> 43.2 | Improved |
| Add probabilistic structural preserving loss (ablation) | PSLR with Lz + Lr · Car196 | R@1 (%): 43.2 -> 43.7 | Improved |
| Increase adaptable factor η from 1 to 10 (ablation) | PSLR · Product | R@1 (%): 40.4 -> 42.3 | Improved |
| Increase adaptable factor η from 1 to 10 (ablation) | PSLR · Product | R@100 (%): 69.7 -> 72.5 | Improved |
| Increase adaptable factor η from 1 to 100 (ablation) | PSLR · CIFAR-10 | kNN accuracy (%): 83.6 -> 85.2 | Improved |
| Optimize latent representations instead of instance features (ablation) | PSLR · Car196 | R@1 (%): 41.3 -> 43.7 | Improved |
| Use a stronger backbone, ResNet50 rather than ResNet18 (ablation) | PSLR · Product | R@1 (%): 52.2 -> 61.6 | Improved |
Closest papers by meaning
- Beyond Supervised vs. Unsupervised: Representative Benchmarking and Analysis of Image Representation Learning (CVPR 2022 · similarity 0.86)
- Simple Unsupervised Graph Representation Learning (AAAI 2022 · similarity 0.86)
- Latent Smooth Skeleton Embedding (AAAI 2017 · similarity 0.86)
- Provable Gaussian Embedding with One Observation (NeurIPS 2018 · similarity 0.86)
- Structural Entropy Guided Probabilistic Coding (AAAI 2025 · similarity 0.85)
- Joint‑Embedding vs Reconstruction: Provable Benefits of Latent Space Prediction for Self‑Supervised Learning (NeurIPS 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.
