Communication-efficient distributed optimization in networks with gradient tracking and variance reduction

Keyphrases

• Gradient Tracking 100%
• Local Averaging 66%
• SARAH 66%
• Approximate Newton 33%
• Global Gradient Estimates 33%
• Multi-agent Learning 33%
• Non-smooth Penalties 33%
• Decentralized Optimization 33%
• Machining Optimization 33%
• Data Homogeneity 33%
• Composite Optimization 33%

Computer Science

• multiagent learning 50%
• Quadratic Loss 50%