Semi-Decentralized Inference in Heterogeneous Graph Neural Networks for Traffic Demand Forecasting: An Edge-Computing Approach

Keyphrases

• Computing Approach 100%
• Edge Computing 100%
• Graph Neural Network 100%
• Decentralized Inference 100%
• Heterogeneous Graph Neural Network 100%
• Traffic Demand Prediction 100%
• LSTM Algorithm 33%
• Large-scale Application 16%
• Prediction Accuracy 16%
• Order of Magnitude 16%
• Local Nodes 16%
• Neural Network Model 16%
• Communication Capabilities 16%
• Inter-node Communication 16%
• Computation Overhead 16%
• Cloudlet 16%
• Communication Overhead 16%
• Service Demand 16%
• Demand Forecasting 16%
• Inference Time 16%
• Graph Structure 16%
• Neural Network Method 16%
• Cellular Base Station 16%
• Provider Profit 16%
• Supply Forecasting 16%
• Computation Capability 16%
• Edge Modes 16%
• Service Supply 16%
• Taxi Service 16%
• Inference Approach 16%
• Network Operation 16%
• Customer Experience 16%
• Node Features 16%
• City Area 16%
• Neural Network Inference 16%
• Inference Scheme 16%

Computer Science

• Edge Computing 100%
• Traffic Demand 100%
• Graph Neural Network 100%
• Long Short-Term Memory Network 25%
• Prediction Accuracy 12%
• Data Center 12%
• Network Inference 12%
• Cloudlet 12%
• Neural Network Model 12%
• Neural Network Approach 12%
• Network Operation 12%
• Customer Experience 12%