DMCA
Large-scale Matrix Factorization with Distributed Stochastic Gradient Descent (2011)
Cached
Download Links
| Venue: | In KDD |
| Citations: | 68 - 7 self |
BibTeX
@INPROCEEDINGS{Gemulla11large-scalematrix,
author = {Rainer Gemulla and Peter J. Haas and Yannis Sismanis and Christina Teflioudi and Faraz Makari},
title = {Large-scale Matrix Factorization with Distributed Stochastic Gradient Descent},
booktitle = {In KDD},
year = {2011},
pages = {69--77}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
We provide a novel algorithm to approximately factor large matrices with millions of rows, millions of columns, and billions of nonzero elements. Our approach rests on stochastic gradient descent (SGD), an iterative stochastic optimization algorithm. Based on a novel “stratified ” variant of SGD, we obtain a new matrixfactorization algorithm, called DSGD, that can be fully distributed and run on web-scale datasets using, e.g., MapReduce. DSGD can handle a wide variety of matrix factorizations and has good scalability properties. 1
Keyphrases
distributed stochastic gradient descent large-scale matrix factorization large matrix novel stratified variant nonzero element stochastic gradient descent matrix factorization wide variety good scalability property novel algorithm web-scale datasets new matrixfactorization algorithm iterative stochastic optimization algorithm
