Research map / Graph learning / Graph algorithms and clustering
ASIL : Augmented Structural Information Learning for Deep Graph Clustering in Hyperbolic Space
Li Sun, Zhenhao Huang, Yujie Wang, Hongbo Lv, Chunyang Liu, Hao Peng, Philip S. Yu
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 |
|---|---|---|---|
| add neural Lorentz boost and augmented structural entropy to LSEnet | LSEnet · Cora | NMI: 63.97 -> 65.78 | Improved |
| add neural Lorentz boost and augmented structural entropy to LSEnet | LSEnet · Citeseer | NMI: 52.26 -> 55.01 | Improved |
| add neural Lorentz boost and augmented structural entropy to LSEnet | LSEnet · AMAP | NMI: 71.72 -> 75.81 | Improved |
| add neural Lorentz boost and augmented structural entropy to LSEnet | LSEnet · Computer | NMI: 55.03 -> 56.31 | Improved |
| add neural Lorentz boost and augmented structural entropy to LSEnet | LSEnet · Pubmed | NMI: 34.25 -> 38.13 | Improved |
| replace heuristic classic structural entropy tree with deep hyperbolic partitioning tree | CSE · Cora | NMI: 60.82 -> 65.78 | Improved |
| replace heuristic classic structural entropy tree with deep hyperbolic partitioning tree | CSE · Citeseer | NMI: 47.12 -> 55.01 | Improved |
| use differentiable level-wise structural entropy with kNN virtual graph (linear complexity) | ASIL vs CSE · Cora | running time: 30x time cost of CSE | Improved |
| use ASIL instead of DGCluster, the fastest baseline | ASIL · Citeseer | running time (s): 79.3 -> 22.1 | Improved |
| use ASIL instead of S3GC | ASIL · Citeseer | running time (s): 1247.2 -> 22.1 | Improved |
| use ASIL instead of Congregate-type best baselines on AMAP NMI | ASIL · AMAP | NMI: 77.12 -> 75.81 | Hurt |
| use ASIL instead of MAGI on Computer ARI | ASIL · Computer | ARI: 46.27 -> 45.77 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 4 on most shared results.
- LSEnet — LSEnet: Lorentz Structural Entropy Neural Network for Deep Graph Clustering (ICML 2024) · better on 10 of 10 shared results
- DinkNet — Dink-Net: Neural Clustering on Large Graphs (ICML 2023) · better on 2 of 2 shared results
- GC-Flow — GC-Flow: A Graph-Based Flow Network for Effective Clustering (ICML 2023) · better on 1 of 1 shared result
- S3GC — S3GC: Scalable Self-Supervised Graph Clustering (NeurIPS 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- LSEnet: Lorentz Structural Entropy Neural Network for Deep Graph Clustering (ICML 2024 · similarity 0.92)
- Hyperbolic Continuous Structural Entropy for Hierarchical Clustering (AAAI 2026 · similarity 0.91)
- Non-Graph Data Clustering via $\mathcal {O}(n)$ Bipartite Graph Convolution (TPAMI 2022 · similarity 0.89)
- THESAURUS: Contrastive Graph Clustering by Swapping Fused Gromov-Wasserstein Couplings (AAAI 2025 · similarity 0.88)
- The Map Equation Goes Neural: Mapping Network Flows with Graph Neural Networks (NeurIPS 2024 · similarity 0.87)
- Structural Entropy Based Graph Structure Learning for Node Classification (AAAI 2024 · 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.
