Research map / Architectures and efficiency / CNN design
Nested Hyperbolic Spaces for Dimensionality Reduction and Hyperbolic NN Design
Check 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 nested hyperbolic dimensionality reduction instead of EPGA | Hyperbolic tree embeddings · balancedtree, L10 to L2 | reconstruction error: \underline{4.01}$\pm$0.76 -> \textbf{3.35}$\pm$0.05 | Helped |
| Use nested hyperbolic dimensionality reduction instead of EPGA | Hyperbolic tree embeddings · unbalanced1, L10 to L2 | reconstruction error: \underline{3.23}$\pm$0.08 -> \textbf{3.10}$\pm$0.01 | Helped |
| Use nested hyperbolic dimensionality reduction instead of EPGA | Hyperbolic tree embeddings · unbalanced2, L10 to L2 | reconstruction error: \underline{3.33}$\pm$0.46 -> \textbf{3.22}$\pm$0.06 | Helped |
| Use nested hyperbolic dimensionality reduction instead of EPGA | Hyperbolic tree embeddings · phylo tree, L10 to L2 | reconstruction error: \underline{25.93}$\pm$0.99 -> \textbf{24.11}$\pm$0.68 | Helped |
| Use nested hyperbolic dimensionality reduction instead of EPGA | Hyperbolic tree embeddings · diseasome, L10 to L2 | reconstruction error: \underline{9.72}$\pm$0.36 -> \textbf{9.18}$\pm$0.10 | Helped |
| Use nested hyperbolic dimensionality reduction instead of EPGA | Hyperbolic tree embeddings · ca-CSphd, L10 to L2 | reconstruction error: \underline{22.98}$\pm$0.23 -> \textbf{22.68}$\pm$0.40 | Helped |
| Use NHGCN instead of H2H-GCN | Hyperbolic graph networks · Disease link prediction | AUC (%): 97.0±0.3 -> 92.8±0.2 | Hurt |
| Use NHGCN instead of HYBONET | Hyperbolic graph networks · Disease node classification | F1 (%): 94.5±0.8 -> 91.7±0.7 | Hurt |
| Use NHGCN instead of HYBONET | Hyperbolic graph networks · Airport link prediction | AUC (%): 97.0±0.2 -> 97.2±0.3 | Helped |
| Use NHGCN instead of HYBONET | Hyperbolic graph networks · Airport node classification | F1 (%): 92.5±0.9 -> 92.4±0.7 | No clear effect |
| Use NHGCN instead of H2H-GCN | Hyperbolic graph networks · PubMed link prediction | AUC (%): 96.9±0.0 -> 96.9±0.1 | No clear effect |
| Use NHGCN instead of HGCN | Hyperbolic graph networks · PubMed node classification | F1 (%): 80.3±0.3 -> 80.5±0.0 | Helped |
Closest papers by meaning
- Random Laplacian Features for Learning with Hyperbolic Space (ICLR 2023 · similarity 0.91)
- Neural networks on Symmetric Spaces of Noncompact Type (ICLR 2025 · similarity 0.88)
- Fully Hyperbolic Convolutional Neural Networks for Computer Vision (ICLR 2024 · similarity 0.88)
- Siegel Neural Networks (NeurIPS 2025 · similarity 0.88)
- Fast Hyperboloid Decision Tree Algorithms (ICLR 2024 · similarity 0.87)
- Hyperbolic Dataset Distillation (NeurIPS 2025 · 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.