Research map / Graph learning / Graph algorithms and clustering
Detection of Model-Based Planted Pseudo-Cliques in Random Dot Product Graphs by the Adjacency Spectral Embedding and the Graph Encoder Embedding
Tong Qi, Vince Lyzinski
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| embed ASE into R^3 instead of R^2 for pseudo-clique of size sqrt(n) (ablation) | ASE · Planted pseudo-clique RDPG, n=1000 | MAP: 0.0463 -> 0.7313 | Improved |
| embed ASE into R^3 instead of R^2 for pseudo-clique of size n^(2/3) (ablation) | ASE · Planted pseudo-clique RDPG, n=1000 | MAP: 0.6856 -> 0.9224 | Improved |
| embed VGAE into R^3 instead of R^2 for pseudo-clique of size sqrt(n) (ablation) | VGAE · Planted pseudo-clique RDPG, n=1000 | MAP: 0.0626 -> 0.0461 | No clear effect |
| embed ASE into R^3 instead of R^2 for pseudo-clique of size 0.2n (ablation) | ASE · Planted pseudo-clique RDPG, n=1000 | MAP: 0.8103 -> 0.9029 | Improved |
| use Leiden CPM partition (more clusters, GEE2) instead of modularity partition (GEE1) for GEE (ablation) | GEE · Planted pseudo-clique RDPG, n=1000, R^3 | MAP: 0.1006 -> 0.1236 | Improved |
| increase planted pseudo-clique size from sqrt(n) to 0.2n (ablation) | GEE1 · Planted pseudo-clique RDPG, n=1000, R^2 | MAP: 0.0370 -> 0.1943 | Improved |
| increase planted pseudo-clique size from sqrt(n) to 0.2n (ablation) | VGAE · Planted pseudo-clique RDPG, n=1000, R^2 | MAP: 0.0626 -> 0.6870 | Improved |
| plant a pseudo-clique instead of a true clique (ablation) | ASE, GEE, VGAE · Unlabeled RDPG | localization | Hurt |
| increase planted clique size from log(n) to 0.2n on EUemail (ablation) | ASE · EUemail | mean embedding distance: 0.0434 -> 14.5977 | Improved |
| increase planted clique size from log(n) to 0.2n on EUemail (ablation) | GEE · EUemail | mean embedding distance: 0.3923 -> 23.0077 | Improved |
| increase planted clique size from log(n) to 0.2n on EUemail (ablation) | VGAE · EUemail | mean embedding distance: 37.2258 -> 43.1421 | Mixed |
| provide additional clean independent network data for embedding comparison | ASE and GEE · Pseudo-clique localization (theory) | asymptotic localization | Improved |
Closest papers by meaning
- Statistical Inference on Random Dot Product Graphs: a Survey (JMLR 2018 · similarity 0.91)
- Maximum Likelihood Embedding of Logistic Random Dot Product Graphs (AAAI 2020 · similarity 0.88)
- One-Hot Graph Encoder Embedding (TPAMI 2022 · similarity 0.87)
- Out-of-sample extension of graph adjacency spectral embedding (ICML 2018 · similarity 0.87)
- Node Similarities under Random Projections: Limits and Pathological Cases (ICLR 2025 · similarity 0.86)
- Information Loss and Disparate Effects in Network Embeddings (NeurIPS 2026 · 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.
