Research map / Graph learning / Graph self-supervised learning
GNN-as-Judge: Unleashing the Power of LLMs for Graph Few-shot Semi-supervised Learning with GNN Feedback
Ruiyao Xu, Kaize Ding
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 pseudo-labeled nodes from GNN-LLM selection (ablation) | GNN-as-Judge · Cora (3-shot) | accuracy: 69.2 -> 77.9 | Improved |
| add the disagreement node set with preference tuning (ablation) | GNN-as-Judge · Cora (3-shot) | accuracy: 71.0 -> 77.9 | Improved |
| use weakly-supervised fine-tuning (instruction + preference tuning) instead of plain instruction tuning (ablation) | GNN-as-Judge · Cora (3-shot) | accuracy: 77.4 -> 77.9 | Improved |
| add pseudo-labeled nodes from GNN-LLM selection (ablation) | GNN-as-Judge · Citeseer (3-shot) | accuracy: 59.1 -> 73.6 | Improved |
| add the disagreement node set with preference tuning (ablation) | GNN-as-Judge · Citeseer (3-shot) | accuracy: 66.8 -> 73.6 | Improved |
| use weakly-supervised fine-tuning (instruction + preference tuning) instead of plain instruction tuning (ablation) | GNN-as-Judge · Citeseer (3-shot) | accuracy: 72.5 -> 73.6 | Improved |
| add pseudo-labeled nodes from GNN-LLM selection (ablation) | GNN-as-Judge · Pubmed (3-shot) | accuracy: 83.0 -> 87.1 | Improved |
| add the disagreement node set with preference tuning (ablation) | GNN-as-Judge · Pubmed (3-shot) | accuracy: 86.0 -> 87.1 | Improved |
| use weakly-supervised fine-tuning (instruction + preference tuning) instead of plain instruction tuning (ablation) | GNN-as-Judge · Pubmed (3-shot) | accuracy: 85.4 -> 87.1 | Improved |
| add pseudo-labeled nodes from GNN-LLM selection (ablation) | GNN-as-Judge · ogbn-arxiv (3-shot) | accuracy: 53.6 -> 62.2 | Improved |
| add the disagreement node set with preference tuning (ablation) | GNN-as-Judge · ogbn-arxiv (3-shot) | accuracy: 59.9 -> 62.2 | Improved |
| use weakly-supervised fine-tuning (instruction + preference tuning) instead of plain instruction tuning (ablation) | GNN-as-Judge · ogbn-arxiv (3-shot) | accuracy: 61.7 -> 62.2 | Improved |
Later papers that compared with this method
1 later accepted paper reports GNN-as-Judge as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Beyond the Golden Teacher: Enhancing Graph Learning through LLM-GNN Co-teaching (NeurIPS 2026 · better on 10 of 10 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
- Label-free Node Classification on Graphs with Large Language Models (LLMs) (ICLR 2024 · similarity 0.92)
- Preference-driven Knowledge Distillation for Few-shot Node Classification (NeurIPS 2025 · similarity 0.91)
- Beyond the Golden Teacher: Enhancing Graph Learning through LLM-GNN Co-teaching (NeurIPS 2026 · similarity 0.91)
- Leveraging Large Language Models for Node Generation in Few-Shot Learning on Text-Attributed Graphs (AAAI 2025 · similarity 0.90)
- LLMs as Zero-shot Graph Learners: Alignment of GNN Representations with LLM Token Embeddings (NeurIPS 2024 · similarity 0.89)
- Harnessing Explanations: LLM-to-LM Interpreter for Enhanced Text-Attributed Graph Representation Learning (ICLR 2024 · similarity 0.89)
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.
