An Advanced Conjugate Gradient Training Algorithm Based on a Modified Secant Equation
Joint Authors
Livieris, Ioannis E.
Pintelas, Panagiotis
Source
Issue
Vol. 2012, Issue 2012 (31 Dec. 2012), pp.1-9, 9 p.
Publisher
Hindawi Publishing Corporation
Publication Date
2011-12-08
Country of Publication
Egypt
No. of Pages
9
Main Subjects
Abstract EN
Conjugate gradient methods constitute excellent neural network training methods characterized by their simplicity, numerical efficiency, and their very low memory requirements.
In this paper, we propose a conjugate gradient neural network training algorithm which guarantees sufficient descent using any line search, avoiding thereby the usually inefficient restarts.
Moreover, it achieves a high-order accuracy in approximating the second-order curvature information of the error surface by utilizing the modified secant condition proposed by Li et al.
(2007).
Under mild conditions, we establish that the proposed method is globally convergent for general functions under the strong Wolfe conditions.
Experimental results provide evidence that our proposed method is preferable and in general superior to the classical conjugate gradient methods and has a potential to significantly enhance the computational efficiency and robustness of the training process.
American Psychological Association (APA)
Livieris, Ioannis E.& Pintelas, Panagiotis. 2011. An Advanced Conjugate Gradient Training Algorithm Based on a Modified Secant Equation. ISRN Artificial Intelligence،Vol. 2012, no. 2012, pp.1-9.
https://search.emarefa.net/detail/BIM-475485
Modern Language Association (MLA)
Livieris, Ioannis E.& Pintelas, Panagiotis. An Advanced Conjugate Gradient Training Algorithm Based on a Modified Secant Equation. ISRN Artificial Intelligence No. 2012 (2012), pp.1-9.
https://search.emarefa.net/detail/BIM-475485
American Medical Association (AMA)
Livieris, Ioannis E.& Pintelas, Panagiotis. An Advanced Conjugate Gradient Training Algorithm Based on a Modified Secant Equation. ISRN Artificial Intelligence. 2011. Vol. 2012, no. 2012, pp.1-9.
https://search.emarefa.net/detail/BIM-475485
Data Type
Journal Articles
Language
English
Notes
Includes bibliographical references
Record ID
BIM-475485