Research map / Learning theory and algorithms / Clustering
Generalised Brown Clustering and Roll-Up Feature Generation
Leon Derczynski, Sean Chester
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 |
|---|---|---|---|
| decouple active set from output classes (a=2560 vs a=|C|) | Brown clustering · CoNLL'03 NER, |C|=10 | F1: 11.31 -> 20.42 | Improved |
| decouple active set (a=2560 vs a=|C|) | Brown clustering · CoNLL'03 NER, |C|=1000 | F1: 46.47 -> 48.23 | Improved |
| roll up Generalised Brown at 2^l clusters instead of shearing, all features | NER CRF · CoNLL'03 NER | F1: 86.81 -> 87.24 | Improved |
| roll up Generalised Brown at 2^l clusters instead of shearing, cluster features only | NER CRF · CoNLL'03 NER | F1: 48.23 -> 51.13 | Improved |
| roll up at the same cluster counts as shearing, all features | NER CRF · CoNLL'03 NER | F1: 86.81 -> 87.14 | Improved |
| roll-up instead of shearing at l=4 | Generalised Brown features · RCV1 | remaining AMI: 0.6735 -> 0.8285 | Improved |
| roll-up instead of shearing at l=20 | Generalised Brown features · RCV1 | remaining AMI: 2.6301 -> 2.6307 | No clear effect |
| larger active set for a very small clustering (|C|<=45) (ablation) | Generalised Brown · Brown corpus subset | AMI | Hurt |
Closest papers by meaning
- Objective-Based Hierarchical Clustering of Deep Embedding Vectors (AAAI 2021 · similarity 0.84)
- Implicit Annealing in Kernel Spaces: A Strongly Consistent Clustering Approach (TPAMI 2022 · similarity 0.83)
- TC-DWA:Text Clustering with Dual Word-Level Augmentation (AAAI 2023 · similarity 0.83)
- Information-Theoretic Generative Clustering of Documents (AAAI 2025 · similarity 0.83)
- Determine-Then-Ensemble: Necessity of Top-k Union for Large Language Model Ensembling (ICLR 2025 · similarity 0.82)
- Explaining Datasets in Words: Statistical Models with Natural Language Parameters (NeurIPS 2024 · similarity 0.82)
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.
