MentorGNN: Deriving Curriculum for Pre-Training GNNs

Keyphrases

• Graph Neural Network 100%
• Graph Signal 83%
• Training Process 16%
• Feature Space 16%
• Human Intervention 16%
• Relational Data 16%
• Diverse Structure 16%
• Generalization Performance 16%
• Generalization Error 16%
• Learning Paradigm 16%
• Label Information 16%
• Mining Communities 16%
• Downstream Applications 16%
• Different Granularity 16%
• Graph Mining 16%
• Hyperparameters 16%
• Target Domain 16%
• Heterogeneous Graph 16%
• Curriculum Learning 16%
• Domain Adaption 16%
• Optimal Hyperparameters 16%
• Pre-training Strategy 16%
• Graph Pre-training 16%

Computer Science

• Graph Neural Network 100%
• Leaning Parameter 33%
• Training Process 16%
• Feature Space 16%
• Human Intervention 16%
• Subgraphs 16%
• Granularity 16%
• Relational Data 16%
• Generalization Error 16%
• Generalization Performance 16%