Node Classification on Products (Accuracy (%))
89.31AccuracyAdvSynGNN
Evaluation results
| Method | Links | |
|---|---|---|
| AdvSynGNN2026.02 | 89.31 | |
| SCL-GNNBackbone=GCN2026.03 | 88.04 | |
| KCoTEvaluation Protocol=TASK EXPERT2026.05 | 86.8 | |
| KCoTEvaluation Protocol=SINGLE FOCUS2026.05 | 86.39 | |
| KCoTEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 86.25 | |
| SCL-GNNBackbone=GAT2026.03 | 85.69 | |
| CANETBackbone=GCN2026.03 | 85.61 | |
| SGFormer2026.02 | 84.75 | |
| LLAGA-HOEvaluation Protocol=SINGLE FOCUS2026.05 | 84.67 | |
| LLAGA-NDEvaluation Protocol=SINGLE FOCUS2026.05 | 84.6 | |
| LLAGA-NDEvaluation Protocol=TASK EXPERT2026.05 | 84.6 | |
| GraphGPTEvaluation Protocol=TASK EXPERT2026.05 | 84.32 | |
| Graphormer2026.02 | 84.18 | |
| LLAGA-HOEvaluation Protocol=TASK EXPERT2026.05 | 84.18 | |
| SRGNNBackbone=GCN2026.03 | 84.16 | |
| GraphGPTEvaluation Protocol=SINGLE FOCUS2026.05 | 84.15 | |
| CANETBackbone=GAT2026.03 | 84.11 | |
| GTGAN2026.02 | 83.95 | |
| GCN2026.02 | 83.9 | |
| TenGAN2026.02 | 83.8 | |
| NodeFormerEvaluation Protocol=SINGLE FOCUS2026.05 | 83.72 | |
| LLAGA-NDEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 83.58 | |
| Att-GAN2026.02 | 83.45 | |
| LLAGA-HOEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 83.32 | |
| GMP-GL2026.02 | 83.1 | |
| GATEvaluation Protocol=SINGLE FOCUS2026.05 | 83.06 | |
| NodeFormerEvaluation Protocol=TASK EXPERT2026.05 | 82.99 | |
| GraphSAGEEvaluation Protocol=SINGLE FOCUS2026.05 | 82.87 | |
| EERMBackbone=GAT2026.03 | 82.68 | |
| DnnGAN2026.02 | 82.67 | |
| GATEvaluation Protocol=TASK EXPERT2026.05 | 82.61 | |
| SAGNEvaluation Protocol=SINGLE FOCUS2026.05 | 82.58 | |
| GraphSAGEEvaluation Protocol=TASK EXPERT2026.05 | 82.5 | |
| EERMBackbone=GCN2026.03 | 82.42 | |
| GraphGPTEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 82.13 | |
| SPGmodel_category=Graph Foundation Models2026.06 | 81.89 | |
| GATEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 81.83 | |
| GraphSAGEEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 81.62 | |
| GraphMAE2model_category=Self-supervised Pre-training2026.06 | 81.07 | |
| GCNEvaluation Protocol=TASK EXPERT2026.05 | 80.88 | |
| GraphGluemodel_category=Graph Foundation Models2026.06 | 80.88 | |
| GCNEvaluation Protocol=SINGLE FOCUS2026.05 | 80.75 | |
| GraphGAN2026.02 | 80.25 | |
| StableGNNBackbone=GCN2026.03 | 80.23 | |
| GCNEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 80.02 | |
| LargeGT2026.02 | 79.81 | |
| GraphAnymodel_category=Graph Foundation Models2026.06 | 79.49 | |
| SAMGPTmodel_category=Graph Foundation Models2026.06 | 79.1 | |
| GRACEmodel_category=Self-supervised Pre-training2026.06 | 77.68 | |
| BGRLmodel_category=Self-supervised Pre-training2026.06 | 77.16 | |
| CleanBackbone=DeepSeek-V32026.03 | 77.1 | |
| RNDBackbone=DeepSeek-V32026.03 | 76.5 | |
| GraphPrompterEvaluation Protocol=Semi-supervised2026.01 | 76.34 | |
| STACKBackbone=DeepSeek-V32026.03 | 75.9 | |
| SGAttackBackbone=DeepSeek-V32026.03 | 75.8 | |
| FLIPBackbone=DeepSeek-V32026.03 | 75.6 | |
| NodeFormerEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 75.55 | |
| SGCEvaluation Protocol=SINGLE FOCUS2026.05 | 75.47 | |
| Our-structBackbone=DeepSeek-V32026.03 | 75.2 | |
| PGDBackbone=DeepSeek-V32026.03 | 75 | |
| GraphSearch-R-72BEvaluation Protocol=Semi-supervised, Model Scale=72B, Search Variant=R2026.01 | 74.6 | |
| GCNEvaluation Protocol=Semi-supervised2026.01 | 74.47 | |
| WTGIABackbone=DeepSeek-V32026.03 | 74.1 | |
| LLaGA-NDEvaluation Protocol=Semi-supervised2026.01 | 73.32 | |
| GraphSearch-F-72BEvaluation Protocol=Semi-supervised, Model Scale=72B, Search Variant=F2026.01 | 73.2 | |
| LLaGA-HOEvaluation Protocol=Semi-supervised2026.01 | 72.76 | |
| SAGEEvaluation Protocol=Semi-supervised2026.01 | 72.35 | |
| GraphSearch-F-32BEvaluation Protocol=Semi-supervised, Model Scale=32B, Search Variant=F2026.01 | 71.7 | |
| RevGATEvaluation Protocol=Semi-supervised2026.01 | 71.45 | |
| GraphSearch-R-32BEvaluation Protocol=Semi-supervised, Model Scale=32B, Search Variant=R2026.01 | 70.8 | |
| DeepSeek-R1-0528Number of candidate categories=47, Zero-shot=true2026.03 | 70.17 | |
| GraphSSRNumber of candidate categories=47, Zero-shot=true2026.03 | 68.49 | |
| Vicuna-7BEvaluation Protocol=CLASSIFICATION EXPERT2026.05 | 67.69 | |
| Qwen3-14B (Thinking)Number of candidate categories=47, Zero-shot=true2026.03 | 66.54 | |
| DeepSeek-R1-distilled-Qwen2.5-14BNumber of candidate categories=47, Zero-shot=true2026.03 | 66.09 | |
| MLPEvaluation Protocol=Semi-supervised2026.01 | 65.36 | |
| Vicuna-7BEvaluation Protocol=TASK EXPERT2026.05 | 65.25 | |
| Our-textBackbone=DeepSeek-V32026.03 | 60.2 | |
| Ministral-3-14B-Reasoning-2512Number of candidate categories=47, Zero-shot=true2026.03 | 59.89 | |
| GCOPEmodel_category=Graph Foundation Models2026.06 | 58.58 | |
| BadGraphBackbone=DeepSeek-V32026.03 | 55.8 | |
| Vicuna-7BEvaluation Protocol=SINGLE FOCUS2026.05 | 55.63 | |
| GraphTranslatorEvaluation Protocol=Semi-supervised2026.01 | 41.32 | |
| TIGmodel_category=Graph Foundation Models2026.06 | 27.31 | |
| Our-structBackbone=Mistral-7B2026.03 | 12.8 | |
| CleanBackbone=Mistral-7B2026.03 | 12.2 | |
| SGAttackBackbone=Mistral-7B2026.03 | 12.2 | |
| PGDBackbone=Mistral-7B2026.03 | 10 | |
| RNDBackbone=Mistral-7B2026.03 | 9.7 | |
| STACKBackbone=Mistral-7B2026.03 | 9.5 | |
| WTGIABackbone=Mistral-7B2026.03 | 9.1 | |
| Our-textBackbone=Mistral-7B2026.03 | 8.6 | |
| FLIPBackbone=Mistral-7B2026.03 | 8.3 | |
| BadGraphBackbone=Mistral-7B2026.03 | 7.8 |