GNSD: A Gradient-Tracking Based Nonconvex Stochastic Algorithm for Decentralized Optimization

Keyphrases

• Stochastic Algorithm 100%
• Non-convexity 100%
• Decentralized Optimization 100%
• Gradient Tracking 100%
• Convergence Rate 33%
• Decentralized Algorithm 16%
• Distributed Systems 16%
• Numerical Experiments 16%
• State-of-the-art Techniques 16%
• Machine Learning Models 16%
• Sqrt 16%
• Nonconvex Optimization Problems 16%
• Centralized Controller 16%
• Single Machine 16%
• Stochastic Gradient Descent 16%
• Computational Resources 16%
• Multiple Parts 16%
• Big Data Era 16%
• Computational Cluster 16%
• Large-scale Dataset 16%
• First-order Stationary Point 16%

Computer Science

• Stochastic Algorithm 100%
• Convergence Rate 100%
• Big Data 50%
• Optimization Problem 50%
• Stationary Point 50%
• Computational Resource 50%
• Single Machine 50%
• Distributed System 50%
• Machine Learning Model 50%
• Stochastic Gradient Descent 50%

Mathematics

• Stochastics 100%
• Stationary Point 14%
• Single Machine 14%
• Step Size 14%
• Convergence Rate 14%
• Numerical Experiment 14%
• Distributed Systems 14%
• Big Data 14%
• Rate of Convergence 14%