Graph Node Classification and Link Prediction on Multimedia
88.86n-F1GraphSAGEco-train
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GraphSAGEco-trainLLM=GPT-4-turbo2024.05 | 88.86 | 70.25 | 62.37 | |
| GraphSAGELLM=GPT-4-turbo2024.05 | 88.63 | 69.65 | 60.36 | |
| DirectLLM=GPT-4-turbo2024.05 | 88.29 | 69.38 | 60.56 | |
| GraphSAGEco-trainLLM=GPT-3.5-turbo2024.05 | 85.51 | 66.56 | 55.91 | |
| GraphSAGELLM=GPT-3.5-turbo2024.05 | 85.29 | 65.8 | 53.55 | |
| DirectLLM=GPT-3.5-turbo2024.05 | 82.85 | 62.07 | 47.96 | |
| GraphSAGEco-trainLLM=Mistral-7B2024.05 | 76.96 | 55.46 | 33.26 | |
| GraphSAGEco-trainLLM=CodeLlama-13B2024.05 | 76.28 | 55.41 | 37.75 | |
| GraphSAGELLM=Mistral-7B2024.05 | 75.51 | 52.94 | 34.29 | |
| GraphSAGELLM=CodeLlama-13B2024.05 | 74.93 | 54.52 | 38.55 | |
| GraphTokenLLM=CodeLlama-13B2024.05 | 74.57 | 47.6 | 35.06 | |
| GraphTokenLLM=Mistral-7B2024.05 | 72.31 | 42.6 | 30.31 | |
| DirectLLM=Mistral-7B2024.05 | 69.83 | 39.85 | 25.05 | |
| DirectLLM=CodeLlama-13B2024.05 | 68.57 | 41.79 | 24.1 | |
| GraphSAGEco-trainLLM=Vicuna-13B2024.05 | 65.89 | 42.18 | 21.58 | |
| GraphTokenLLM=Vicuna-13B2024.05 | 65.4 | 36.38 | 19.87 | |
| TAPELLM=GPT-4-turbo2024.05 | 63.94 | — | 4.02 | |
| GraphSAGELLM=Vicuna-13B2024.05 | 63.71 | 39.88 | 21.37 | |
| TAPELLM=GPT-3.5-turbo2024.05 | 62.43 | — | 3.87 | |
| TAPELLM=Mistral-7B2024.05 | 58.92 | — | 3.29 | |
| TAPELLM=CodeLlama-13B2024.05 | 58.27 | — | 2.01 | |
| DirectLLM=Vicuna-13B2024.05 | 55.47 | 23.19 | 11.2 | |
| TAPELLM=Vicuna-13B2024.05 | 53.57 | — | 2.07 |