CANITA: Faster Rates for Distributed Convex Optimization with Communication Compression

Keyphrases

• Fast Rates 100%
• Federated Learning 100%
• Accelerated Gradient Method 100%
• Distributed Convex Optimization 100%
• Compressed Communication 100%
• Communication Round 100%
• Communication Compression 100%
• Convex Problem 50%
• Convergence Rate 50%
• Distributed Learning 50%
• Communication Cost 50%
• Gradient Method 50%
• Learning Styles 50%
• Fast Convergence 50%
• Largest Mean 50%
• Distributed Optimization 50%
• Smoothing Parameter 50%
• Target Error 50%
• Fast Convergence Rate 50%
• Compression Parameters 50%
• DIANA 50%
• ANITA 50%

Computer Science

• Fast Convergence 100%
• Federated Learning 100%
• Convex Optimization 100%
• Gradient Descent 50%
• Communication Cost 50%
• Gradient Method 50%
• Distributed Learning 50%
• Distributed Optimization 50%
• Convergence Rate 50%