Research map / Learning theory and algorithms / Clustering
The Catalog Problem: Clustering and Ordering Variable-Sized Sets
Mateusz Maria Jurewicz, Graham W. Taylor, Leon Derczynski
Read the paperCheck 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 cluster-level cardinality prediction module (NOC2) with its loss to the neural clustering pipeline | NOC · Synthetic Catalogs (n=3,4,5) | Compositional score: +11.96, +13.01, +15.65 points over second-best | Improved |
| Use NOC (joint clustering, cardinality prediction, pointer ordering) instead of S2S-based baselines | NOC · PROCAT | V-Measure and Kendall's τ: +11.46 V-Measure and +12.28 Kendall's τ over second-best | Improved |
| Use NOC instead of baselines on the 2D Gaussians task | NOC · 2D Gaussians | V-Measure and Kendall's τ: +1.18 V-Measure over ACP; +2.51 Kendall's τ over S2S-B | Improved |
| Use NOC instead of S2S-based baselines on procedural catalogs | NOC · Synthetic Catalogs structural score | Structural score: +4.54, +3.59, +3.61 over S2S-C/S2S-B | Improved |
| Neural clustering methods with set-to-sequence pointer ordering (NCP/CCP/ACP) vs S2S-B/S2S-C on PROCAT | Neural clustering baselines · PROCAT | Kendall's τ: S2S models outperform NCP, CCP, ACP on ordering by +12.66 points | Hurt |
Closest papers by meaning
- PROCAT: Product Catalogue Dataset for Implicit Clustering, Permutation Learning and Structure Prediction (NeurIPS 2021 · similarity 0.86)
- Unsupervised Order Learning (ICLR 2024 · similarity 0.85)
- Cluster-Wise Hierarchical Generative Model for Deep Amortized Clustering (CVPR 2021 · similarity 0.83)
- Discovering Data Structures: Nearest Neighbor Search and Beyond (NeurIPS 2025 · similarity 0.82)
- Deep Taxonomic Networks for Unsupervised Hierarchical Prototype Discovery (NeurIPS 2025 · similarity 0.81)
- Learning-Augmented Hierarchical Clustering (ICML 2025 · similarity 0.81)
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.
