Accelerated Randomized Coordinate Descent Algorithms for Stochastic\n Optimization and Online Learning

2018/06/05 by Akshita Bhandari, Bhandari, Akshita, Chandramani Singh +1
Computer Science · Decision Sciences · Engineering · #Advanced Bandit Algorithms Research #Error Correcting Code Techniques #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Sparse and Compressive Sensing Techniques #Stochastic Gradient Optimization Techniques

paper · pdf · doi:10.48550/arxiv.1806.01600

openalex publication_date 2018/06/05 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We propose accelerated randomized coordinate descent algorithms for\nstochastic optimization and online learning. Our algorithms have significantly\nless per-iteration complexity than the known accelerated gradient algorithms.\nThe proposed algorithms for online learning have better regret performance than\nthe known randomized online coordinate descent algorithms. Furthermore, the\nproposed algorithms for stochastic optimization exhibit as good convergence\nrates as the best known randomized coordinate descent algorithms. We also show\nsimulation results to demonstrate performance of the proposed algorithms.\n

Citations

• A Stochastic Gradient Method with an Exponential Convergence Rate for Finite Training Sets
• Randomized Block Coordinate Descent for Online and Stochastic Optimization
• Federated Optimization: Distributed Machine Learning for On-Device Intelligence

Related

• Randomized Block Coordinate Descent for Online and Stochastic Optimization
• Even Faster Accelerated Coordinate Descent Using Non-Uniform Sampling
• About accelerated randomized methods
• An Analysis of Asynchronous Stochastic Accelerated Coordinate Descent
• Stability and Generalization for Randomized Coordinate Descent