Research map / Graph learning / GNN architectures
DiP-G: Discrete Prompting for Graph Neural Networks
Yumeng Zhao, Huiying Hu, Shuo Wen, Junjie Shen, Bei Hua
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 |
|---|---|---|---|
| use perturbed Top-k discrete prompts trained end to end instead of soft prompts with test-time Top-k (ablation) | DiP-G · Cora 5-shot (GraphCL) | accuracy: 61.54 -> 65.12 | Improved |
| use perturbed Top-k discrete prompts trained end to end instead of soft prompts with test-time Top-k (ablation) | DiP-G · PubMed 5-shot (GraphCL) | accuracy: 67.84 -> 70.45 | Improved |
| use perturbed Top-k discrete prompts trained end to end instead of soft prompts with test-time Top-k (ablation) | DiP-G · Amazon 5-shot (GraphCL) | accuracy: 25.47 -> 29.25 | Improved |
| use perturbed Top-k discrete prompts trained end to end instead of soft prompts with test-time Top-k (ablation) | DiP-G · Minesweeper 5-shot (GraphCL) | accuracy: 68.10 -> 70.84 | Improved |
| use I-MLE gradient estimator instead of straight-through estimator (ablation) | DiP-G · Cora 5-shot (GraphCL) | accuracy: 63.84 -> 65.12 | Improved |
| use I-MLE gradient estimator instead of straight-through estimator (ablation) | DiP-G · PubMed 5-shot (GraphCL) | accuracy: 69.13 -> 70.45 | Improved |
| use I-MLE gradient estimator instead of straight-through estimator (ablation) | DiP-G · Amazon 5-shot (GraphCL) | accuracy: 27.80 -> 29.25 | Improved |
| use I-MLE gradient estimator instead of straight-through estimator (ablation) | DiP-G · Minesweeper 5-shot (GraphCL) | accuracy: 69.53 -> 70.84 | Improved |
| increase sparsity k from small values to 8 | DiP-G · Cora and PubMed 5-shot (GraphCL) | accuracy | Improved |
| increase sparsity k beyond 8 | DiP-G · Cora and PubMed 5-shot (GraphCL) | accuracy | No clear effect |
| increase candidate budget m up to 128 then beyond | DiP-G · Cora and PubMed 5-shot (GraphCL) | accuracy | Improved |
| increase screening buffer size r | DiP-G screened target solve · Cora and PubMed | fallback rate | Improved |
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.
- GraphTOP — GraphTOP: Graph Topology-Oriented Prompting for Graph Neural Networks (NeurIPS 2025) · better on 7 of 7 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
- GraphTOP: Graph Topology-Oriented Prompting for Graph Neural Networks (NeurIPS 2025 · similarity 0.93)
- Edge Prompt Tuning for Graph Neural Networks (ICLR 2025 · similarity 0.91)
- ProG: A Graph Prompt Learning Benchmark (NeurIPS 2024 · similarity 0.90)
- One Prompt Fits All: Universal Graph Adaptation for Pretrained Models (NeurIPS 2025 · similarity 0.90)
- HeterGP: Bridging Heterogeneity in Graph Neural Networks with Multi-View Prompting (AAAI 2025 · similarity 0.89)
- Spectral Reversal: Counteracting Singular Value Bias for Graph Prompting (NeurIPS 2026 · 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.
