Research map / Architectures and efficiency / RNNs and normalization
Using k-Way Co-Occurrences for Learning Word Embeddings
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 |
|---|---|---|---|
| include co-occurrences up to k=3 (ablation) | k-way embeddings · MEN | Spearman correlation: 70.42 -> 70.92 | Helped |
| include co-occurrences up to k=3 (ablation) | k-way embeddings · DV | accuracy: 75.34 -> 75.96 | Helped |
| include co-occurrences up to k=3 (ablation) | k-way embeddings · CR | accuracy: 79.19 -> 79.52 | Helped |
| include co-occurrences up to k=5 (ablation) | k-way embeddings · WS | Spearman correlation: 59.68 -> 60.54 | Helped |
| include co-occurrences up to k=5 (ablation) | k-way embeddings · SUBJ | accuracy: 82.20 -> 82.60 | Helped |
| include co-occurrences up to k=3 (ablation) | k-way embeddings · TR | accuracy: 72.43 -> 72.92 | Helped |
| include co-occurrences up to k=3 (ablation) | k-way embeddings · MC | Spearman correlation: 79.17 -> 79.92 | Helped |
| include co-occurrences up to k=3 (ablation) | k-way embeddings · RG | Spearman correlation: 78.63 -> 77.51 | No clear effect |
| include co-occurrences up to k=3 (ablation) | k-way embeddings · SL | Spearman correlation: 34.76 -> 34.65 | Hurt |
| include co-occurrences beyond k=2 (ablation) | k-way embeddings · short-text classification datasets | accuracy | Helped |
Closest papers by meaning
- Semantic Structure-Based Word Embedding by Incorporating Concept Convergence and Word Divergence (AAAI 2018 · similarity 0.88)
- Joint Word Representation Learning Using a Corpus and a Semantic Lexicon (AAAI 2016 · similarity 0.87)
- Symmetries in language statistics shape the geometry of model representations (ICML 2026 · similarity 0.87)
- On the Emergence of Linear Analogies in Word Embeddings (NeurIPS 2025 · similarity 0.87)
- By Tying Embeddings You Are Assuming the Distributional Hypothesis (ICML 2024 · similarity 0.86)
- Co-occurrence is not Factual Association in Language Models (NeurIPS 2024 · similarity 0.85)
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.