|
|
Bimonthly Since 1986 |
ISSN 1004-9037
|
|
|
|
|
Publication Details |
Edited by: Editorial Board of Journal of Data Acquisition and Processing
P.O. Box 2704, Beijing 100190, P.R. China
Sponsored by: Institute of Computing Technology, CAS & China Computer Federation
Undertaken by: Institute of Computing Technology, CAS
Published by: SCIENCE PRESS, BEIJING, CHINA
Distributed by:
China: All Local Post Offices
|
|
|
|
|
|
|
|
|
|
|
05 July-September 2023, Volume 38 Issue 4
|
|
|
|
OPTIMIZATION ALGORITHM
S. Selvakumari #1, Dr. M. Durairaj., #2
Journal of Data Acquisition and Processing, 2023, 38 (4): 1992-1997 .
|
Abstract
Preparing the deep learning models includes learning of the boundaries to meet the goal work. Normally the objective is to limit the misfortune brought about during the learning measure. In a regulated method of learning, a model is given the information tests and their particular results. At the point when a model produces a yield, it contrasts it and the ideal yield and afterward takes the distinction of produced and wanted yields and afterward endeavors to carry the produced yield near the wanted yield. This is accomplished through streamlining calculations. An enhancement calculation goes through a few cycles until intermingling to work on the precision of the model. There are a few sorts of streamlining techniques created to address the challenges related with the learning system. Six of these have been taken up to be inspected in this review to acquire bits of knowledge about their complexities. The techniques examined are SGD, nesterov force, rmsprop, adam, adagrad, adadelta.
Keyword
SGD, Momentum, RMSPROP.
PDF Download (click here)
|
|
|
|
|