Research map / Learning theory and algorithms / Clustering
Objective-Based Hierarchical Clustering of Deep Embedding Vectors
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 |
|---|---|---|---|
| Use B++&C instead of BKMeans | hierarchical clustering of deep embeddings · ImageNet, ImageNetV2, NaBirds, Twitter, Wikipedia, SST-2 | normalized CKMM: .23 -> .30; .70 -> .83; .61 -> .71; .11 -> .60; .23 -> .58; .15 -> .49 | Helped |
| Use B++&C instead of the strongest listed baseline | hierarchical clustering of deep embeddings · ImageNet, ImageNetV2, NaBirds, Twitter, Wikipedia, SST-2 | normalized MW: .37 -> .40; .51 -> .51; .71 -> .73; .27 -> .44; .40 -> .40; .45 -> .51 | Helped |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · Twitter | normalized CKMM: .12 -> .10 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · Wikipedia | normalized CKMM: .30 -> .17 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · NaBirds | normalized CKMM: .71 -> .57 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · ImageNetV2 | normalized CKMM: .83 -> .26 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · ImageNet | normalized CKMM: .30 -> .23 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · SST-2 | normalized CKMM: .49 -> .46 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · Twitter | normalized MW: .44 -> .23 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · NaBirds | normalized MW: .73 -> .67 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · SST-2 | normalized MW: .51 -> .46 | Hurt |
| Set the bisection imbalance parameter to δ=0 (ablation) | B++&C · ImageNet | normalized MW: .40 -> .37 | Hurt |
Closest papers by meaning
- From Logits to Hierarchies: Hierarchical Clustering made Simple (ICML 2025 · similarity 0.91)
- Image Clustering via the Principle of Rate Reduction in the Age of Pretrained Models (ICLR 2024 · similarity 0.87)
- Image Clustering with External Guidance (ICML 2024 · similarity 0.87)
- Image Clustering Conditioned on Text Criteria (ICLR 2024 · similarity 0.87)
- Language-Assisted Image Clustering Guided by Discriminative Relational Signals and Adaptive Semantic Centers (NeurIPS 2026 · similarity 0.86)
- Deep Taxonomic Networks for Unsupervised Hierarchical Prototype Discovery (NeurIPS 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.