Research map / Graph learning / Graph algorithms and clustering
Hyperbolic Entailment Cones for Learning Hierarchical 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 |
|---|---|---|---|
| Use hyperbolic entailment cones instead of Order Emb | WordNet link prediction, 5 and 10 dimensions · 0% non-basic edges | test F1 (%): 34.4% -> 29.2%; 43.0% -> 32.2% | Hurt |
| Use hyperbolic entailment cones instead of the strongest baseline | WordNet link prediction, 5 and 10 dimensions · 10%, 25%, and 50% non-basic edges | test F1 (%): 71.3% -> 80.1%; 75.4% -> 85.9%; 78.2% -> 86.0%; 82.0% -> 91.0%; 83.6% -> 92.8%; 85.3% -> 94.4% | Helped |
| Initialize Euclidean cones with Poincaré embeddings rather than Simple Euclidean embeddings (ablation) | Our Euclidean Cones · WordNet link prediction | test F1 | No clear effect |
| Use full Riemannian optimization rather than the retraction approximation (ablation) | Hyperbolic models · WordNet link prediction | test F1 | No clear effect |
| Add transitive-closure edges to training | Our Hyperbolic Cones · WordNet link prediction | test F1: 80.1% -> 86.0% -> 92.8%; 85.9% -> 91.0% -> 94.4% | Helped |
Closest papers by meaning
- Shadow Cones: A Generalized Framework for Partial Order Embeddings (ICLR 2024 · similarity 0.91)
- Weighted Embeddings for Low-Dimensional Graph Representation (AAAI 2025 · similarity 0.89)
- Hyperbolic Embeddings of Supervised Models (NeurIPS 2024 · similarity 0.89)
- Low-distortion and GPU-compatible Tree Embeddings in Hyperbolic Space (ICML 2025 · similarity 0.88)
- Learning Representations for Hierarchies with Minimal Support (NeurIPS 2024 · similarity 0.87)
- Compositional Entailment Learning for Hyperbolic Vision-Language Models (ICLR 2025 · similarity 0.86)
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.