An Accelerated Gradient Method for Trace Norm Minimization
J. Ji. (2009)An optimization method for matrix factorization type problems that directly minimizes the trace norm. Needs a full SVD of U and M at each step thus not applicable to large scale problems..
An optimization method for matrix factorization type problems that directly minimizes the trace norm. Needs a full SVD of U and M at each step thus not applicable to large scale problems.
Please log in to take part in the discussion (add own reviews or comments).
Cite this publication
More citation styles
- please select -
%0 Conference Paper
%1 ShuJie09
%A Ji, Jieping Ye Shuiwang
%D 2009
%E 2009, ICML
%K factorization optimization
%T An Accelerated Gradient Method for Trace Norm Minimization
%U http://www.cs.mcgill.ca/~icml2009/papers/151.pdf
@inproceedings{ShuJie09,
added-at = {2009-07-20T16:14:37.000+0200},
author = {Ji, Jieping Ye Shuiwang},
biburl = {https://www.bibsonomy.org/bibtex/2863ba37ca3d933a885c74440dab04fa7/alexis99},
description = {Optimization for Multiple learning, and matrix factorization type problems, positive: seems fast negative: requires svd in each step thus limited in terms of data sizes.},
editor = {2009, ICML},
interhash = {11ec570d2274d161862b8326e019f63b},
intrahash = {863ba37ca3d933a885c74440dab04fa7},
keywords = {factorization optimization},
note = {An optimization method for matrix factorization type problems that directly minimizes the trace norm. Needs a full SVD of U and M at each step thus not applicable to large scale problems.},
timestamp = {2009-07-20T16:14:37.000+0200},
title = {An Accelerated Gradient Method for Trace Norm Minimization},
url = {http://www.cs.mcgill.ca/~icml2009/papers/151.pdf},
year = 2009
}