Research map / Label-efficient and robust learning / Noisy labels
Partially View-Aligned Representation Learning With Noise-Robust Contrastive Loss
Check 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 the noise-robust contrastive loss instead of using vanilla contrastive loss (ablation) | MvCLN · NoisyMNIST | ACC: 88.2 -> 92.17 | Helped |
| Add the noise-robust contrastive loss instead of using vanilla contrastive loss (ablation) | MvCLN · NoisyMNIST | ARI: 75.89 -> 84.51 | Helped |
| Use more randomly sampled negative pairs per positive pair (ablation) | MvCLN · NoisyMNIST | clustering performance | Helped |
| Increase the available aligned-data proportion (ablation) | MvCLN · NoisyMNIST | clustering performance | Helped |
| Switch from Stage 1 to Stage 2 at a mean negative-pair distance between 0.2m and 1.0m (ablation) | MvCLN · NoisyMNIST | clustering performance | No clear effect |
| Omit the initial vanilla-contrastive optimization stage (ablation) | MvCLN · NoisyMNIST | clustering performance | Hurt |
| Use category-level alignment with MvCLN rather than the PVC baseline | MvCLN · Caltech-101 (partially aligned) | ARI: 17.98 -> 51.44 | Helped |
| Use category-level alignment with MvCLN rather than the PVC baseline | MvCLN · Reuters (partially aligned) | ACC: 42.07 -> 56.62 | Helped |
Closest papers by meaning
- Robust Contrastive Learning Against Noisy Views (CVPR 2022 · similarity 0.89)
- ROLL: Robust Noisy Pseudo-label Learning for Multi-View Clustering with Noisy Correspondence (CVPR 2025 · similarity 0.89)
- Your contrastive learning problem is secretly a distribution alignment problem (NeurIPS 2024 · similarity 0.87)
- $\mathbb{X}$-Sample Contrastive Loss: Improving Contrastive Learning with Sample Similarity Graphs (ICLR 2025 · similarity 0.86)
- On the Similarities of Embeddings in Contrastive Learning (ICML 2025 · similarity 0.86)
- CFDM: Contrastive Fusion and Disambiguation for Multi-View Partial-Label Learning (AAAI 2025 · 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.