AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

Learning on Large-scale Text-attributed Graphs via Variational Inference

Jianan Zhao, Meng Qu, Chaozhuo Li, Hao Yan, Qian Liu, Rui Li, Xing Xie, Jian Tang

ICLR 2023 · Graph learning › Graph self-supervised learning · approach: LLMs and graph foundation models · 25 citations (OpenAlex)

Read the paper

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
Use GLEM-GNN instead of RevGAT with GIANT featuresRevGAT on ogbn-arxiv · ogbn-arxiv testaccuracy: 75.90 ± 0.19 -> 76.97 ± 0.19Improved
Use GLEM-LM instead of fine-tuning the LM aloneDeBERTa on ogbn-arxiv · ogbn-arxiv testaccuracy: 74.13 ± 0.04 -> 75.71 ± 0.24Improved
Use GLEM-GNN instead of SAGN+ with GIANT featuresSAGN+ on ogbn-products · ogbn-products testaccuracy: 86.67 ± 0.09 -> 87.36 ± 0.07Improved
Use GLEM-LM instead of fine-tuning the LM aloneDeBERTa on ogbn-products · ogbn-products testaccuracy: 79.63 ± 0.12 -> 84.83 ± 0.04Improved
Use GLEM-GNN instead of GAMLP+ with GIANT featuresGAMLP+ on ogbn-papers100M · ogbn-papers100M testaccuracy: 69.67 ± 0.05 -> 70.36 ± 0.02Improved
Use GLEM-LM instead of fine-tuning the LM aloneDeBERTa on ogbn-papers100M · ogbn-papers100M testaccuracy: 63.52 ± 0.06 -> 66.71 ± 0.25Improved
Add GNN pseudo-label guidance to LM training instead of using gold labels only (α=0) (ablation)GLEM-LM on ogbn-arxiv · ogbn-arxivaccuracyImproved
Vary the GNN pseudo-label weight (ablation)GLEM-GNN on ogbn-arxiv · ogbn-arxivaccuracyNo 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