Research map / Graph learning / Graph self-supervised learning
Latent Graph Inference using Product Manifolds
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 a product manifold rather than the Euclidean dDGM embedding (ablation) | GCN latent graph inference · Cora | accuracy (%): 82.11 -> 86.63 | Helped |
| Use a product manifold rather than the Euclidean dDGM embedding (ablation) | GCN latent graph inference · CiteSeer | accuracy (%): 72.35 -> 75.42 | Helped |
| Use a product manifold rather than the Euclidean dDGM embedding (ablation) | GCN latent graph inference · Physics | accuracy (%): 95.96 -> 96.21 | Helped |
| Use a product manifold rather than a single model space (ablation) | GCN latent graph inference · homophilic and heterophilic benchmarks | accuracy (%) | Mixed |
| Use a hyperbolic single-space embedding instead of a Euclidean embedding (ablation) | GCN latent graph inference · Texas | accuracy (%): 80.00 -> 79.44 | Hurt |
| Use a hyperbolic single-space embedding instead of a Euclidean embedding (ablation) | GCN latent graph inference · Squirrel | accuracy (%): 34.35 -> 35.00 | Helped |
| Use a spherical-spherical product manifold (ablation) | GCN latent graph inference · Cora | accuracy (%): 82.11 -> 65.96 | Hurt |
| Use a spherical-spherical product manifold (ablation) | GCN latent graph inference · PubMed | accuracy (%): 87.69 -> 87.82 | Helped |
| Infer the latent graph with dDGM | GCN on Texas · Texas | accuracy (%): 41.66 -> 81.67 | Helped |
| Use a latent graph rather than the input graph | GCN on Wisconsin · Wisconsin | accuracy (%): 47.20 -> 89.33 | Helped |
| Use symbolic rather than dense distance matrices | GCN-dDGM-E · Cora | runtime (s): 2.98 -> 2.64 | Helped |
| Use symbolic rather than dense distance matrices | GCN-dDGM-E · CiteSeer | runtime (s): 3.50 -> 2.73 | Helped |
Closest papers by meaning
- From Latent Graph to Latent Topology Inference: Differentiable Cell Complex Module (ICLR 2024 · similarity 0.90)
- Neural Snowflakes: Universal Latent Graph Inference via Trainable Latent Geometries (ICLR 2024 · similarity 0.89)
- Neural Latent Geometry Search: Product Manifold Inference via Gromov-Hausdorff-Informed Bayesian Optimization (NeurIPS 2023 · similarity 0.88)
- Mixed-Curvature Geometric Latent Diffusion Model for Graph Generation (NeurIPS 2026 · similarity 0.88)
- Learning Latent Graph Structures and their Uncertainty (ICML 2025 · similarity 0.88)
- Graph Geometry-Preserving Autoencoders (ICML 2024 · 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.