Research map / Label-efficient and robust learning / Recognition and long-tail
Semantic Structure-Based Word Embedding by Incorporating Concept Convergence and Word Divergence
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 |
|---|---|---|---|
| add concept convergence and word divergence structure losses from WordNet | CBOW · Word similarity WS | Spearman correlation: 0.638 -> 0.688 | Helped |
| add semantic-structure losses from WordNet | Skip-gram · Word similarity WS | Spearman correlation: 0.631 -> 0.694 | Helped |
| add semantic-structure losses | CBOW · Word similarity VERB | Spearman correlation: 0.402 -> 0.402 | No clear effect |
| add semantic-structure losses | Skip-gram · Word analogy total | accuracy: 56.0 -> 57.2 | Helped |
| add semantic-structure losses | CBOW · Word analogy total | accuracy: 48.7 -> 49.9 | Helped |
| add semantic-structure losses | CBOW · 20NewsGroup | accuracy: 78.1 -> 81.4 | Helped |
| add semantic-structure losses | Skip-gram · 20NewsGroup | accuracy: 80.2 -> 81.7 | Helped |
| add semantic-structure losses | CBOW · RCV1 query expansion | MAP: 44.3 -> 45.1 | Helped |
| use Retrofit post-hoc refinement instead of joint structure learning | CBOW/Skip-gram · 20NewsGroup | accuracy: 78.1 -> 75.6 (CBOW); 80.2 -> 77.4 (Skip-gram) | Hurt |
| use Retrofit refinement | CBOW · Word analogy total | accuracy: 48.7 -> 39.1 Tot (CBOW) | Hurt |
| use Retrofit refinement | CBOW/Skip-gram · Word similarity | Spearman correlation | Mixed |
| increase negative samples beyond 5 | SENSE · WordSim353 | Spearman correlation | Hurt |
Closest papers by meaning
- Understanding the Semantic Content of Sparse Word Embeddings Using a Commonsense Knowledge Base (AAAI 2020 · similarity 0.92)
- Joint Word Representation Learning Using a Corpus and a Semantic Lexicon (AAAI 2016 · similarity 0.89)
- Improving Word Embeddings with Convolutional Feature Learning and Subword Information (AAAI 2017 · similarity 0.89)
- Unseen Word Representation by Aligning Heterogeneous Lexical Semantic Spaces (AAAI 2019 · similarity 0.89)
- On the Emergence of Linear Analogies in Word Embeddings (NeurIPS 2025 · similarity 0.88)
- By Tying Embeddings You Are Assuming the Distributional Hypothesis (ICML 2024 · similarity 0.87)
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.