Parallel and Distributed Computation: Numerical Methods, 1989. ,
A Parallel Mixture of SVMs for Very Large Scale Problems, Neural Computation, vol.20, issue.5 ,
DOI : 10.1162/neco.1991.3.1.79
Taming the wild: a unified analysis of Hogwild!-style algorithms, NIPS, 2015. ,
SAGA: A fast incremental gradient method with support for non-strongly convex composite objectives, NIPS, 2014. ,
URL : https://hal.archives-ouvertes.fr/hal-01016843
Asynchronous stochastic convex optimization, NIPS, 2015. ,
Variance reduced stochastic gradient descent with neighbors, NIPS, 2015. ,
URL : https://hal.archives-ouvertes.fr/hal-01248672
PASSCoDe: Parallel ASynchronous Stochastic dual Co-ordinate Descent, ICML, 2015. ,
Accelerating stochastic gradient descent using predictive variance reduction, NIPS, 2013. ,
A stochastic gradient method with an exponential convergence rate for finite training sets, NIPS, 2012. ,
URL : https://hal.archives-ouvertes.fr/hal-00674995
RCV1: A new benchmark collection for text categorization research, JMLR, 2004. ,
Asynchronous parallel stochastic gradient for nonconvex optimization, NIPS, 2015. ,
An asynchronous parallel stochastic coordinate descent algorithm, Journal of Machine Learning Research, vol.16, pp.285-322, 2015. ,
Identifying suspicious URLs, Proceedings of the 26th Annual International Conference on Machine Learning, ICML '09, 2009. ,
DOI : 10.1145/1553374.1553462
Perturbed iterate analysis for asynchronous stochastic optimization, 2015. ,
Hogwild: a lock-free approach to parallelizing stochastic gradient descent, NIPS, 2011. ,
On variance reduction in stochastic gradient descent and its asynchronous variants, NIPS, 2015. ,
Minimizing finite sums with the stochastic average gradient, 2013. ,
URL : https://hal.archives-ouvertes.fr/hal-00860051
Stochastic dual coordinate ascent methods for regularized loss, JMLR, vol.14, issue.1, pp.567-599, 2013. ,
Fast asynchronous parallel stochastic gradient descent, AAAI, 2016. ,
On our third dataset, the associated task is a binary classification problem (down from 7 classes originally, following the pre-treatment of Collobert et al. [2]). The features are cartographic variables ,
We only use our fourth dataset for non-parallel experiments and a specific compare-and-swap test. It constitutes of UseNet articles taken from four discussion groups (simulated auto racing, simulated aviation, real autos, real aviation) ,
All experiments were run on a Dell PowerEdge 920 machine with 4 Intel Xeon E7-4830v2 processors with 10 2 ,
All algorithms were implemented in the Scala language and the software stack consisted of a Linux operating system running Scala 2 ,