Research map / Label-efficient and robust learning / Recognition and long-tail
Geometry of Polysemy
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 5 instead of 2 clusters (ablation) | K-Grassmeans · SemEval-2010 | V-Measure (x100): 7.30 -> 14.50 | Helped |
| Use 5 instead of 2 clusters (ablation) | K-Grassmeans · SemEval-2010 | Paired F-score (x100): 57.14 -> 44.07 | Hurt |
| Use 5 instead of 2 clusters (ablation) | K-Grassmeans · Make-Sense-2016 | V-Measure (x100): 29.30 -> 34.40 | Helped |
| Use 5 instead of 2 clusters (ablation) | K-Grassmeans · Make-Sense-2016 | Paired F-score (x100): 64.58 -> 58.17 | Hurt |
| Use 5 instead of 2 clusters (ablation) | K-Grassmeans · SCWS | Spearman correlation (x100), HardSim: 58.40 -> 53.14 | Hurt |
| Use 5 instead of 2 clusters (ablation) | K-Grassmeans · SCWS-lite | Spearman correlation (x100), HardSim: 4.97 -> 22.28 | Helped |
| Use 5 instead of 2 clusters (ablation) | K-Grassmeans · SCWS-lite | Spearman correlation (x100), SoftSim: 5.07 -> 16.54 | Helped |
| Use soft rather than hard decoding (ablation) | K-Grassmeans (#cluster=2) · SCWS | Spearman correlation (x100): 58.40 -> 63.67 | Helped |
| Use soft rather than hard decoding (ablation) | K-Grassmeans (#cluster=5) · SCWS-lite | Spearman correlation (x100): 22.28 -> 17.83 | Hurt |
| Use K-Grassmeans instead of the strongest reported baseline | K-Grassmeans (#cluster=2) · SemEval-2010 | Paired F-score (x100): 48.43 -> 57.14 | Helped |
| Use K-Grassmeans instead of Huang 2012 | K-Grassmeans (#cluster=2) · Make-Sense-2016 | Paired F-score (x100): 15.9 -> 64.58 | Helped |
| Use K-Grassmeans lexeme representations instead of MSSG | SCWS-lite · SCWS-lite | Spearman correlation (x100), HardSim: 20.67 -> 22.28 | Helped |
Closest papers by meaning
- Geometric Relationship between Word and Context Representations (AAAI 2018 · similarity 0.90)
- SensEmBERT: Context-Enhanced Sense Embeddings for Multilingual Word Sense Disambiguation (AAAI 2020 · similarity 0.87)
- FIRE: Semantic Field of Words Represented as Non-Linear Functions (NeurIPS 2022 · similarity 0.86)
- Leveraging Lexical Substitutes for Unsupervised Word Sense Induction (AAAI 2018 · similarity 0.86)
- Joint Word Representation Learning Using a Corpus and a Semantic Lexicon (AAAI 2016 · similarity 0.86)
- The Geometry of Categorical and Hierarchical Concepts in Large Language Models (ICLR 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.