Research map / Label-efficient and robust learning / Self-supervised and contrastive
Contextually Affinitive Neighborhood Refinery for Deep Clustering
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 |
|---|---|---|---|
| replace local Euclidean neighborhood with ConAff neighborhood (ablation) | BYOL-based clustering · CIFAR-10 | ACC: 89.6 -> 91.1 | Helped |
| add online boundary detection to local neighborhood clustering (ablation) | BYOL-based clustering · CIFAR-10 | ACC: +1.9% | Helped |
| add online boundary detection to ConAff neighborhood clustering (ablation) | BYOL-based clustering · CIFAR-10 | ACC: +1.4% | Helped |
| add progressive relaxation to boundary filtering (ablation) | ConAff neighborhood with online boundary detection · CIFAR-10 | ACC: +0.7% | Helped |
| use progressively relaxed boundary filtering with local neighborhood (ablation) | BYOL-based clustering · CIFAR-10 | ACC: 89.6 -> 91.6 | No clear effect |
| set initial filtering fraction to fr=1, removing boundary filtering (ablation) | CoNR · CIFAR-10 | clustering performance | Hurt |
| add ConNR* neighborhood module | SimSiam · CIFAR-10 | ACC: +4.8% | Helped |
| add ConNR* neighborhood module | MoCo v2 · CIFAR-10 | ACC: +11.3% | Helped |
| use ConNR instead of BYOL | BYOL · STL-10 | ACC: +10.1% | Helped |
| use ConNR instead of BYOL | BYOL · ImageNet-Dogs | ACC: +6.2% | Helped |
| use ConNR rather than ProPos | deep clustering · CIFAR-10 | ACC: +1.6% | Helped |
| use ConNR rather than ProPos | deep clustering · CIFAR-20 | ACC: +2.6% | Helped |
Closest papers by meaning
- Cluster-Aware Similarity Diffusion for Instance Retrieval (ICML 2024 · similarity 0.86)
- Clustering Properties of Self-Supervised Learning (ICML 2025 · similarity 0.85)
- Breaking the Reclustering Barrier in Centroid-based Deep Clustering (ICLR 2025 · similarity 0.85)
- Image Clustering with External Guidance (ICML 2024 · similarity 0.85)
- Hard Regularization to Prevent Deep Online Clustering Collapse without Data Augmentation (AAAI 2024 · similarity 0.84)
- You Can Trust Your Clustering Model: A Parameter-free Self-Boosting Plug-in for Deep Clustering (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.