Optimizing Machine Learning on Apache Spark in HPC Environments

Keyphrases

• Machine Learning 100%
• High Performance Computing 100%
• Computing Environment 100%
• Apache Spark 100%
• Asynchronous Stochastic Gradient Descent 60%
• Stochastic Gradient Algorithm 40%
• Allreduce 40%
• Training Data 20%
• Large-scale Cluster 20%
• Task-oriented 20%
• Statistical Techniques 20%
• Full Potential 20%
• Non-blocking 20%
• Decision Algorithm 20%
• Convergence Rate 20%
• Cluster Computing 20%
• Decision Model 20%
• Data Analytics 20%
• Compute-intensive 20%
• Management Communication 20%
• High-performance Computing Cluster 20%
• Learning Analytics 20%
• Communication Cost 20%
• Memory Consumption 20%
• Deep Neural Network 20%
• NVIDIA 20%
• Synchronization Overheads 20%
• ImageNet Dataset 20%
• Communication Synchronization 20%
• Memory Management 20%
• Distributed Framework 20%
• Computer Neural Networks 20%
• Commodity Servers 20%
• Data Flow Framework 20%
• Heterogeneous Cluster 20%
• Graphics Card 20%
• Accelerator Card 20%
• GoogLeNet Model 20%
• Cost Overhead 20%
• Multiple Machines 20%
• Distributed Dataflows 20%
• Distributed Data Representation 20%
• Cluster of Nodes 20%
• Synchronous Functions 20%

Computer Science

• Machine Learning 100%
• Learning System 100%
• High-Performance Computing 100%
• Computing Environment 100%
• Apache Spark 100%
• Gradient Descent 60%
• Training Data 20%
• Decision-Making 20%
• Neural Network Model 20%
• Statistical Technique 20%
• Data Representation 20%
• Nonblocking 20%
• Computer Cluster 20%
• Data Analytics 20%
• Convergence Rate 20%
• Communication Cost 20%
• Heterogeneous Cluster 20%
• Multiple Machine 20%
• Memory Management 20%
• Intensive Process 20%
• Deep Neural Network 20%