Presentation	2012-11-08 Optimization of Feature Allocation on Parallel Stochastic Gradient Descent for Sparse Data Kohei HAYASHI, Ryohei FUJIMAKI,
PDF Download Page	PDF download Page Link
Abstract(in Japanese)	(See Japanese page)
Abstract(in English)	In feature-wise parallelization of stochastic gradient descent, feature allocation governs the total computational time, which does not scale with the number of parallel units as we expected in general. We tackle the problem and propose efficient feature allocation methods. We evaluate the performance of them both theoretically and empirically with synthetic data.
Keyword(in Japanese)	(See Japanese page)
Keyword(in English)	Stochastic gradient descent / parallel computing / discrete optimization
Paper #	IBISML2012-77
Date of Issue

Conference Information
Committee	IBISML
Conference Date	2012/10/31(1days)
Place (in Japanese)	(See Japanese page)
Place (in English)
Topics (in Japanese)	(See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To	Information-Based Induction Sciences and Machine Learning (IBISML)
Language	JPN
Title (in Japanese)	(See Japanese page)
Sub Title (in Japanese)	(See Japanese page)
Title (in English)	Optimization of Feature Allocation on Parallel Stochastic Gradient Descent for Sparse Data
Sub Title (in English)
Keyword(1)	Stochastic gradient descent
Keyword(2)	parallel computing
Keyword(3)	discrete optimization
1st Author's Name	Kohei HAYASHI
1st Author's Affiliation	Graduate School of Information Science and Technology, University of Tokyo:JSPS()
2nd Author's Name	Ryohei FUJIMAKI
2nd Author's Affiliation	NEC Laboratories America
Date	2012-11-08
Paper #	IBISML2012-77
Volume (vol)	vol.112
Number (no)	279
Page	pp.pp.-
#Pages	8
Date of Issue