AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Graph is a Natural Regularization: Revisiting Vector Quantization for Graph Representation Learning

Zian Zhai, Fan Li, Xingyu Tan, Xiaoyang Wang, Wenjie Zhang

ICML 2026 · Graph learning › Graph self-supervised learning · approach: Semi-supervised node representation learning

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
add RGVQ (Gumbel-Softmax plus structure-aware regularization) to Graph VQGFT · Cora codebook perplexity (K=512)perplexity: 94.47 -> 211.69Improved
add RGVQGFT · PubMed codebook perplexity (K=512)perplexity: 4.14 -> 319.09Improved
add RGVQGFT · Average over 7 pretrain-finetune datasetsaccuracy / ROC-AUC: 79.50 -> 81.42Improved
add RGVQGQT · Cora node classificationaccuracy: 86.44 -> 88.34Improved
add RGVQGQT · Roman node classificationaccuracy: 89.85 -> 90.98Improved
add SimVQ collapse mitigationGFT · Average over 7 datasetsaccuracy / ROC-AUC: 79.50 -> 75.97Hurt
add EMA codebook updateGFT · Average over 7 datasetsaccuracy / ROC-AUC: 79.50 -> 79.73No clear effect
add reparameterization with structural regularization only (remove feature signal) (ablation)RGVQ · Coraaccuracy: 78.35 -> 79.87Improved
add both structural and feature regularization versus structure only (ablation)RGVQ · PubMedaccuracy: 76.32 -> 77.46Improved
remove Gumbel-Softmax reparameterization (ablation)RGVQ · Cora/PubMed/WikiCSperplexity: 319.09 -> 4.14 (PubMed)Hurt
lower Gumbel-Softmax temperature (ablation)RGVQ · codebook perplexityperplexityImproved
increase number of contrastive samples (ablation)RGVQ · codebook perplexityperplexityNo clear effect

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