STEM: A Stochastic Two-Sided Momentum Algorithm Achieving Near-Optimal Sample and Communication Complexities for Federated Learning.
Prashant KhanduriPranay SharmaHaibo YangMingyi HongJia LiuKetan RajawatPramod K. VarshneyPublished in: CoRR (2021)
Keyphrases
- learning algorithm
- preprocessing
- worst case
- optimal solution
- monte carlo
- optimization algorithm
- theoretical analysis
- recognition algorithm
- dynamic programming
- np hard
- cost function
- active learning
- prior knowledge
- incremental learning
- learning process
- high accuracy
- learning problems
- linear programming
- learning speed
- objective function
- detection algorithm
- convergence rate
- matching algorithm
- stochastic approximation
- online learning
- simulated annealing
- supervised learning
- computational cost
- experimental evaluation
- search space
- reinforcement learning