Decentralized Local Updates with Dual-Slow Estimation and Momentum-Based Variance-Reduction for Non-Convex Optimization.
Kangyang LuoKunkun ZhangShengbo ZhangXiang LiMing GaoPublished in: ECAI (2023)
Keyphrases
- convex optimization
- variance reduction
- primal dual
- gradient estimation
- importance sampling
- monte carlo
- sample size
- convex optimization problems
- markov chain
- total variation
- objective function
- parameter estimation
- confidence intervals
- naive bayes classifier
- estimation error
- natural images
- particle filter
- semi supervised
- knn
- training set