Research map / Label-efficient and robust learning / Recognition and long-tail
Two Knowledge-based Methods for High-Performance Sense Distribution Learning
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 entropy-filtered Personalized PageRank sense distributions (EnDi) instead of LexSemTM | MFS sense prediction · Senseval/SemEval all-words | F1: 0.49 -> 0.66 | Helped |
| Use domain-aware distributions (DaD) instead of LexSemTM | MFS sense prediction · Senseval/SemEval all-words | F1: 0.49 -> 0.61 | Helped |
| Replace manual SemCor MFS with learned EnDi distributions | MFS baseline · Senseval/SemEval all-words | F1: 0.68 -> 0.66 | Hurt |
| Use degree-based distributions instead of learned ones | sense distribution · Senseval/SemEval all-words | F1: 0.66 -> 0.43 | Hurt |
| Use EnDi instead of SemCor distribution | sense distribution · Bennett et al. 50-lemma gold | JSD: 0.255 -> 0.099 | Helped |
| Use DaD instead of BFS baseline | MFS for Italian · SemEval-2015 Italian | F1: 0.52 -> 0.61 | Helped |
| Use DaD instead of BFS baseline | MFS for Spanish · SemEval-2015 Spanish | F1: 0.53 -> 0.58 | Helped |
| Use EnDi instead of BFS baseline | MFS for Spanish · SemEval-2015 Spanish | F1: 0.53 -> 0.52 | No clear effect |
| Use domain-specific corpus (DaD) instead of WordNet MFS | domain MFS · SemEval-2013 domains | F1: 0.61 -> 0.79 | Helped |
| Use EnDi instead of WordNet MFS on domains | domain MFS · SemEval-2013/2015 domains | F1: 0.46 vs 0.50 (Medicine) | Mixed |
| Entropy filtering of sentence-level distributions (EnDi) vs unfiltered domain method on JSD | EnDi · SemCor similarity (sys) | JSD: 0.12 -> 0.06 | Helped |
Closest papers by meaning
- Knowledge-based Word Sense Disambiguation using Topic Models (AAAI 2018 · similarity 0.89)
- XL-WSD: An Extra-Large and Cross-Lingual Evaluation Framework for Word Sense Disambiguation (AAAI 2021 · similarity 0.88)
- SensEmBERT: Context-Enhanced Sense Embeddings for Multilingual Word Sense Disambiguation (AAAI 2020 · similarity 0.88)
- CSI: A Coarse Sense Inventory for 85% Word Sense Disambiguation (AAAI 2020 · similarity 0.87)
- AutoSense Model for Word Sense Induction (AAAI 2019 · similarity 0.87)
- Beyond the Seen: Bounded Distribution Estimation for Open-Vocabulary Learning (NeurIPS 2025 · similarity 0.83)
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.