AI Paper Atlas

Research map / Graph learning / Graph algorithms and clustering

Anchor-Driven Nyström for Deep Graph-Level Clustering

Jiaxin Wang, Wenxuan Tu, Lingren Wang, Jieren Cheng, Yue Yang

AAAI 2026 · Graph learning › Graph algorithms and clustering · approach: Neural and local graph clustering · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
make Nystrom anchors learnable instead of fixed (ablation)ANGC · MUTAGclustering ACC (%): 74.54 -> 85.18Improved
make Nystrom anchors learnable instead of fixed (ablation)ANGC · DDclustering ACC (%): 56.40 -> 75.27Improved
make Nystrom anchors learnable instead of fixed (ablation)ANGC · COLLABclustering ACC (%): 55.42 -> 64.21Improved
make Nystrom anchors learnable instead of fixed (ablation)ANGC · NCI1clustering ACC (%): 53.39 -> 61.48Improved
add anchor response discrepancy regularization (ablation)ANGC · MUTAGclustering ACC (%): 81.53 -> 85.18Improved
add anchor response discrepancy regularization (ablation)ANGC · COLLABclustering ACC (%): 58.86 -> 64.21Improved
add anchor response discrepancy regularization (ablation)ANGC · PTC-MRclustering ACC (%): 56.81 -> 58.26Improved
add anchor response discrepancy regularization (ablation)ANGC · PTC-MRNMI (%): 2.73 -> 2.65No clear effect
use ANGC instead of best baseline GNN/kernel clusteringANGC · PTC-MRclustering ACC (%): 59.98 -> 58.26Hurt
use ANGC instead of GPC/DCGLCANGC · DDclustering ACC (%): 69.52 -> 75.27Improved
use ANGC kernel anchors vs DCGLC (runtime)ANGC · REDDIT-M-12Kone-epoch runtime (s): 1069.05 -> 559.14Improved
use ANGC kernel anchors vs GraphCL (runtime)ANGC · MUTAGone-epoch runtime (s): 3.14 -> 4.27Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy