Abstract in

DMCA

Abstract in (1993)

Cached

Download Links

by Cheng-tao Chu , Yuanyuan Yu , Sang Kyun Kim , Gary Bradski , Kunle Olukotun , Yi-an Lin , Andrew Y. Ng

Venue:	I 13-GHz f~ graded-base SiGe HBTs,” 51st Device Res. Conf
Citations:	1 - 0 self

BibTeX

@INPROCEEDINGS{Chu93abstractin,
    author = {Cheng-tao Chu and Yuanyuan Yu and Sang Kyun Kim and Gary Bradski and Kunle Olukotun and Yi-an Lin and Andrew Y. Ng},
    title = {Abstract in},
    booktitle = {I 13-GHz f~ graded-base SiGe HBTs,” 51st Device Res. Conf},
    year = {1993},
    pages = {2100}
}

OpenURL

Abstract

We are at the beginning of the multicore era. Computers will have increasingly many cores (processors), but there is still no good programming framework for these architectures, and thus no simple and unified way for machine learning to take advantage of the potential speed up. In this paper, we develop a broadly applicable parallel programming method, one that is easily applied to many different learning algorithms. Our work is in distinct contrast to the tradition in machine learning of designing (often ingenious) ways to speed up a single algorithm at a time. Specifically, we show that algorithms that fit the Statistical Query model [15] can be written in a certain “summation form, ” which allows them to be easily parallelized on multicore computers. We adapt Google’s map-reduce [7] paradigm to demonstrate this parallel speed up technique on a variety of learning algorithms including locally weighted linear regression (LWLR), k-means, logistic regression

Keyphrases

machine learning potential speed linear regression good programming framework distinct contrast single algorithm certain summation form multicore era many core unified way logistic regression multicore computer applicable parallel programming method statistical query model