Gossip Dual Averaging for Decentralized Optimization of Pairwise Functions - Archive ouverte HAL Access content directly
Conference Papers Year :

Gossip Dual Averaging for Decentralized Optimization of Pairwise Functions

(1) , (2) , (1) , (1)
1
2

Abstract

In decentralized networks (of sensors, connected objects, etc.), there is an important need for efficient algorithms to optimize a global cost function , for instance to learn a global model from the local data collected by each computing unit. In this paper, we address the problem of decentralized minimization of pairwise functions of the data points, where these points are distributed over the nodes of a graph defining the communication topology of the network. This general problem finds applications in ranking, distance metric learning and graph inference, among others. We propose new gossip algorithms based on dual averaging which aims at solving such problems both in synchronous and asynchronous settings. The proposed framework is flexible enough to deal with constrained and regularized variants of the optimization problem. Our theoretical analysis reveals that the proposed algorithms preserve the convergence rate of centralized dual averaging up to an additive bias term. We present numerical simulations on Area Under the ROC Curve (AUC) maximization and metric learning problems which illustrate the practical interest of our approach.
Fichier principal
Vignette du fichier
icml16.pdf (458.03 Ko) Télécharger le fichier
Vignette du fichier
icml16_supp.pdf (600.66 Ko) Télécharger le fichier
Origin : Files produced by the author(s)
Origin : Files produced by the author(s)
Loading...

Dates and versions

hal-01329315 , version 1 (09-06-2016)

Identifiers

  • HAL Id : hal-01329315 , version 1

Cite

Igor Colin, Aurélien Bellet, Joseph Salmon, Stéphan Clémençon. Gossip Dual Averaging for Decentralized Optimization of Pairwise Functions. International Conference on Machine Learning (ICML 2016), Jun 2016, New York, United States. ⟨hal-01329315⟩
232 View
81 Download

Share

Gmail Facebook Twitter LinkedIn More