English

Federated Empirical Risk Minimization via Second-Order Method

Machine Learning 2023-05-30 v1 Distributed, Parallel, and Cluster Computing

Abstract

Many convex optimization problems with important applications in machine learning are formulated as empirical risk minimization (ERM). There are several examples: linear and logistic regression, LASSO, kernel regression, quantile regression, pp-norm regression, support vector machines (SVM), and mean-field variational inference. To improve data privacy, federated learning is proposed in machine learning as a framework for training deep learning models on the network edge without sharing data between participating nodes. In this work, we present an interior point method (IPM) to solve a general ERM problem under the federated learning setting. We show that the communication complexity of each iteration of our IPM is O~(d3/2)\tilde{O}(d^{3/2}), where dd is the dimension (i.e., number of features) of the dataset.

Keywords

Cite

@article{arxiv.2305.17482,
  title  = {Federated Empirical Risk Minimization via Second-Order Method},
  author = {Song Bian and Zhao Song and Junze Yin},
  journal= {arXiv preprint arXiv:2305.17482},
  year   = {2023}
}
R2 v1 2026-06-28T10:48:22.048Z