New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val)

Kotaro Hirasawa, Kazuyuki Togo, Junichi Murata, Masanao Ohbayashi, Ning Shao, Takayuki Furuzuki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, a new random search method Ras Val for Neural Networks (NN) learning is proposed. RasVal (Random Search with Variable Search Length) is a kind of random search and it can find a global minimum instead of a local minimum using the capability of intensified and diversified searches. The main different point of RasVal from commonly used Random Search Methods (RSM) is that the shape of the probability density function for random searching can be adjusted based on the information of success or failure of the search. First, RasVal is described and after that, performance between RasVal, Back Propagation Method (BP) and Back Propagation Method with momentum (Mom. BP) are compared. The performance is evaluated by the simulations which include both static and dynamic Neural Networks (NN) learning problems. In the simulations, NN is trained to realize nonlinear functions and to control a nonlinear crane system by using RasVal, BP and Mom. BP. Simulation results show that Ras Val is superior or nearly equal to BP and Mom. BP because of the ability of intensification and diversification of the search.

Original languageEnglish
Title of host publicationIEEE International Conference on Neural Networks - Conference Proceedings
Editors Anon
Place of PublicationPiscataway, NJ, United States
PublisherIEEE
Pages1602-1607
Number of pages6
Volume2
Publication statusPublished - 1998
Externally publishedYes
EventProceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3) - Anchorage, AK, USA
Duration: 1998 May 41998 May 9

Other

OtherProceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3)
CityAnchorage, AK, USA
Period98/5/498/5/9

Fingerprint

Backpropagation
Neural networks
Cranes
Probability density function
Momentum

ASJC Scopus subject areas

  • Software

Cite this

Hirasawa, K., Togo, K., Murata, J., Ohbayashi, M., Shao, N., & Furuzuki, T. (1998). New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val). In Anon (Ed.), IEEE International Conference on Neural Networks - Conference Proceedings (Vol. 2, pp. 1602-1607). Piscataway, NJ, United States: IEEE.

New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val). / Hirasawa, Kotaro; Togo, Kazuyuki; Murata, Junichi; Ohbayashi, Masanao; Shao, Ning; Furuzuki, Takayuki.

IEEE International Conference on Neural Networks - Conference Proceedings. ed. / Anon. Vol. 2 Piscataway, NJ, United States : IEEE, 1998. p. 1602-1607.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hirasawa, K, Togo, K, Murata, J, Ohbayashi, M, Shao, N & Furuzuki, T 1998, New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val). in Anon (ed.), IEEE International Conference on Neural Networks - Conference Proceedings. vol. 2, IEEE, Piscataway, NJ, United States, pp. 1602-1607, Proceedings of the 1998 IEEE International Joint Conference on Neural Networks. Part 1 (of 3), Anchorage, AK, USA, 98/5/4.
Hirasawa K, Togo K, Murata J, Ohbayashi M, Shao N, Furuzuki T. New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val). In Anon, editor, IEEE International Conference on Neural Networks - Conference Proceedings. Vol. 2. Piscataway, NJ, United States: IEEE. 1998. p. 1602-1607
Hirasawa, Kotaro ; Togo, Kazuyuki ; Murata, Junichi ; Ohbayashi, Masanao ; Shao, Ning ; Furuzuki, Takayuki. / New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val). IEEE International Conference on Neural Networks - Conference Proceedings. editor / Anon. Vol. 2 Piscataway, NJ, United States : IEEE, 1998. pp. 1602-1607
@inproceedings{88f8e03f13a043f3a8e5bae8d08734e3,
title = "New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val)",
abstract = "In this paper, a new random search method Ras Val for Neural Networks (NN) learning is proposed. RasVal (Random Search with Variable Search Length) is a kind of random search and it can find a global minimum instead of a local minimum using the capability of intensified and diversified searches. The main different point of RasVal from commonly used Random Search Methods (RSM) is that the shape of the probability density function for random searching can be adjusted based on the information of success or failure of the search. First, RasVal is described and after that, performance between RasVal, Back Propagation Method (BP) and Back Propagation Method with momentum (Mom. BP) are compared. The performance is evaluated by the simulations which include both static and dynamic Neural Networks (NN) learning problems. In the simulations, NN is trained to realize nonlinear functions and to control a nonlinear crane system by using RasVal, BP and Mom. BP. Simulation results show that Ras Val is superior or nearly equal to BP and Mom. BP because of the ability of intensification and diversification of the search.",
author = "Kotaro Hirasawa and Kazuyuki Togo and Junichi Murata and Masanao Ohbayashi and Ning Shao and Takayuki Furuzuki",
year = "1998",
language = "English",
volume = "2",
pages = "1602--1607",
editor = "Anon",
booktitle = "IEEE International Conference on Neural Networks - Conference Proceedings",
publisher = "IEEE",

}

TY - GEN

T1 - New random search method for Neural Networks learning - Random Search with Variable Search Length (Ras Val)

AU - Hirasawa, Kotaro

AU - Togo, Kazuyuki

AU - Murata, Junichi

AU - Ohbayashi, Masanao

AU - Shao, Ning

AU - Furuzuki, Takayuki

PY - 1998

Y1 - 1998

N2 - In this paper, a new random search method Ras Val for Neural Networks (NN) learning is proposed. RasVal (Random Search with Variable Search Length) is a kind of random search and it can find a global minimum instead of a local minimum using the capability of intensified and diversified searches. The main different point of RasVal from commonly used Random Search Methods (RSM) is that the shape of the probability density function for random searching can be adjusted based on the information of success or failure of the search. First, RasVal is described and after that, performance between RasVal, Back Propagation Method (BP) and Back Propagation Method with momentum (Mom. BP) are compared. The performance is evaluated by the simulations which include both static and dynamic Neural Networks (NN) learning problems. In the simulations, NN is trained to realize nonlinear functions and to control a nonlinear crane system by using RasVal, BP and Mom. BP. Simulation results show that Ras Val is superior or nearly equal to BP and Mom. BP because of the ability of intensification and diversification of the search.

AB - In this paper, a new random search method Ras Val for Neural Networks (NN) learning is proposed. RasVal (Random Search with Variable Search Length) is a kind of random search and it can find a global minimum instead of a local minimum using the capability of intensified and diversified searches. The main different point of RasVal from commonly used Random Search Methods (RSM) is that the shape of the probability density function for random searching can be adjusted based on the information of success or failure of the search. First, RasVal is described and after that, performance between RasVal, Back Propagation Method (BP) and Back Propagation Method with momentum (Mom. BP) are compared. The performance is evaluated by the simulations which include both static and dynamic Neural Networks (NN) learning problems. In the simulations, NN is trained to realize nonlinear functions and to control a nonlinear crane system by using RasVal, BP and Mom. BP. Simulation results show that Ras Val is superior or nearly equal to BP and Mom. BP because of the ability of intensification and diversification of the search.

UR - http://www.scopus.com/inward/record.url?scp=0031619892&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0031619892&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:0031619892

VL - 2

SP - 1602

EP - 1607

BT - IEEE International Conference on Neural Networks - Conference Proceedings

A2 - Anon, null

PB - IEEE

CY - Piscataway, NJ, United States

ER -