Analyzing weight distribution of neural networks

Jinwook Go, Chulhee Lee

Research output: Contribution to conferencePaper

2 Citations (Scopus)

Abstract

In this paper, we investigate weight distribution of neural networks in order to understand and improve training process of neural networks. Generally, it takes a long time to train neural networks. However, when a new problem is presented, neural networks have to be trained again. No matter how many times neural networks have been trained before, the neural networks must be trained again without any benefit from the previous training process. The training process can be viewed as finding a solution point in the weight space. In this paper, we investigate the distribution of the solution points in the weight space to understand and speed up the training process.

Original languageEnglish
Pages1154-1157
Number of pages4
Publication statusPublished - 1999 Dec 1
EventInternational Joint Conference on Neural Networks (IJCNN'99) - Washington, DC, USA
Duration: 1999 Jul 101999 Jul 16

Other

OtherInternational Joint Conference on Neural Networks (IJCNN'99)
CityWashington, DC, USA
Period99/7/1099/7/16

Fingerprint

Neural networks

All Science Journal Classification (ASJC) codes

  • Software

Cite this

Go, J., & Lee, C. (1999). Analyzing weight distribution of neural networks. 1154-1157. Paper presented at International Joint Conference on Neural Networks (IJCNN'99), Washington, DC, USA, .
Go, Jinwook ; Lee, Chulhee. / Analyzing weight distribution of neural networks. Paper presented at International Joint Conference on Neural Networks (IJCNN'99), Washington, DC, USA, .4 p.
@conference{5aed4151cc8f457eb41d27a76b560c1e,
title = "Analyzing weight distribution of neural networks",
abstract = "In this paper, we investigate weight distribution of neural networks in order to understand and improve training process of neural networks. Generally, it takes a long time to train neural networks. However, when a new problem is presented, neural networks have to be trained again. No matter how many times neural networks have been trained before, the neural networks must be trained again without any benefit from the previous training process. The training process can be viewed as finding a solution point in the weight space. In this paper, we investigate the distribution of the solution points in the weight space to understand and speed up the training process.",
author = "Jinwook Go and Chulhee Lee",
year = "1999",
month = "12",
day = "1",
language = "English",
pages = "1154--1157",
note = "International Joint Conference on Neural Networks (IJCNN'99) ; Conference date: 10-07-1999 Through 16-07-1999",

}

Go, J & Lee, C 1999, 'Analyzing weight distribution of neural networks' Paper presented at International Joint Conference on Neural Networks (IJCNN'99), Washington, DC, USA, 99/7/10 - 99/7/16, pp. 1154-1157.

Analyzing weight distribution of neural networks. / Go, Jinwook; Lee, Chulhee.

1999. 1154-1157 Paper presented at International Joint Conference on Neural Networks (IJCNN'99), Washington, DC, USA, .

Research output: Contribution to conferencePaper

TY - CONF

T1 - Analyzing weight distribution of neural networks

AU - Go, Jinwook

AU - Lee, Chulhee

PY - 1999/12/1

Y1 - 1999/12/1

N2 - In this paper, we investigate weight distribution of neural networks in order to understand and improve training process of neural networks. Generally, it takes a long time to train neural networks. However, when a new problem is presented, neural networks have to be trained again. No matter how many times neural networks have been trained before, the neural networks must be trained again without any benefit from the previous training process. The training process can be viewed as finding a solution point in the weight space. In this paper, we investigate the distribution of the solution points in the weight space to understand and speed up the training process.

AB - In this paper, we investigate weight distribution of neural networks in order to understand and improve training process of neural networks. Generally, it takes a long time to train neural networks. However, when a new problem is presented, neural networks have to be trained again. No matter how many times neural networks have been trained before, the neural networks must be trained again without any benefit from the previous training process. The training process can be viewed as finding a solution point in the weight space. In this paper, we investigate the distribution of the solution points in the weight space to understand and speed up the training process.

UR - http://www.scopus.com/inward/record.url?scp=0033346033&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0033346033&partnerID=8YFLogxK

M3 - Paper

SP - 1154

EP - 1157

ER -

Go J, Lee C. Analyzing weight distribution of neural networks. 1999. Paper presented at International Joint Conference on Neural Networks (IJCNN'99), Washington, DC, USA, .