DistDGL: Distributed Graph Neural Network Training for Billion-Scale Graphs

Keyphrases

• Neural Network Training 100%
• Graph Neural Network 100%
• Billion-scale 100%
• Distributed Graph Neural Networks 100%
• Mini 66%
• Network Development 33%
• Model Accuracy 33%
• Halo 33%
• Design Decisions 33%
• Development Framework 33%
• Neural Network Model 33%
• Network Form 33%
• Communication Overhead 33%
• Fraud Detection 33%
• Quality Model 33%
• Sparse Embedding 33%
• Ego Networks 33%
• Graph Structure 33%
• Parallel Efficiency 33%
• Training Approaches 33%
• Linear Speedup 33%
• Distributed Computation 33%
• Balance Constraints 33%
• Local Nodes 33%
• Min-cut 33%
• Memory Scalability 33%
• Transductive 33%
• Cut Graph 33%
• Deep Graph Library 33%
• Synchronous Training 33%
• Parallel Memory 33%
• Training Epochs 33%
• Popular 33%
• Graph Partitioning Algorithm 33%

Computer Science

• Graph Neural Network 100%
• Neural Network Training 100%
• Neural Network Model 25%
• Communication Overhead 25%
• Model Accuracy 25%
• Linear Speedup 25%
• Development Framework 25%
• Distributed Computation 25%
• Parallel Efficiency 25%
• Graph Partition 25%
• quality model 25%
• Structured Data 25%
• Fraud Detection 25%