A Stochastic Large-scale Machine Learning Algorithm for Distributed Features and Observations
This work addresses the challenge of handling large-scale data sets that exceed single-machine capacities, offering a solution for distributed machine learning practitioners, though it appears incremental as it builds on existing distributed optimization methods.
The authors tackled the problem of distributed optimization in machine learning when both features and observations are distributed across multiple computers, proposing a stochastic algorithm that demonstrates superior performance in early iterations compared to a benchmark, with computational experiments in Spark showing this advantage.
As the size of modern data sets exceeds the disk and memory capacities of a single computer, machine learning practitioners have resorted to parallel and distributed computing. Given that optimization is one of the pillars of machine learning and predictive modeling, distributed optimization methods have recently garnered ample attention, in particular when either observations or features are distributed, but not both. We propose a general stochastic algorithm where observations, features, and gradient components can be sampled in a double distributed setting, i.e., with both features and observations distributed. Very technical analyses establish convergence properties of the algorithm under different conditions on the learning rate (diminishing to zero or constant). Computational experiments in Spark demonstrate a superior performance of our algorithm versus a benchmark in early iterations of the algorithm, which is due to the stochastic components of the algorithm.