Opening the black box -Data driven visualization of neural networks

Fan Yin Tzeng, Kwan-Liu Ma

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

44 Citations (Scopus)

Abstract

Artificial neural networks are computer software or hardware models inspired by the structure and behavior of neurons in the human nervous system. As a powerful learning tool, increasingly neural networks have been adopted by many large-scale information processing applications but there is no a set of well defined criteria for choosing a neural network. The user mostly treats a neural network as a black box and cannot explain how learning from input data was done nor how performance can be consistently ensured. We have experimented with several information visualization designs aiming to open the black box to possibly uncover underlying dependencies between the input data and the output data of a neural network. In this paper, we present our designs and show that the visualizations not only help us design more efficient neural networks, but also assist us in the process of using neural networks for problem solving such as performing a classification task.

Original languageEnglish (US)
Title of host publicationVIS 05
Subtitle of host publicationIEEE Visualization 2005, Proceedings
Number of pages1
DOIs
StatePublished - Dec 1 2005
EventVIS 05: IEEE Visualization 2005, Proceedings - Minneapolis, MN, United States
Duration: Oct 23 2005Oct 28 2005

Other

OtherVIS 05: IEEE Visualization 2005, Proceedings
CountryUnited States
CityMinneapolis, MN
Period10/23/0510/28/05

Fingerprint

Data visualization
Neural networks
Visualization
Neurology
Computer hardware
Neurons

Keywords

  • Artificial Neural Network
  • Classification
  • Information Visualization
  • Machine Learning
  • Visualization Application

ASJC Scopus subject areas

  • Software
  • Computer Science(all)
  • Engineering(all)
  • Computer Graphics and Computer-Aided Design

Cite this

Tzeng, F. Y., & Ma, K-L. (2005). Opening the black box -Data driven visualization of neural networks. In VIS 05: IEEE Visualization 2005, Proceedings [1566031] https://doi.org/10.1109/VIS.2005.73

Opening the black box -Data driven visualization of neural networks. / Tzeng, Fan Yin; Ma, Kwan-Liu.

VIS 05: IEEE Visualization 2005, Proceedings. 2005. 1566031.

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

Tzeng, FY & Ma, K-L 2005, Opening the black box -Data driven visualization of neural networks. in VIS 05: IEEE Visualization 2005, Proceedings., 1566031, VIS 05: IEEE Visualization 2005, Proceedings, Minneapolis, MN, United States, 10/23/05. https://doi.org/10.1109/VIS.2005.73
Tzeng FY, Ma K-L. Opening the black box -Data driven visualization of neural networks. In VIS 05: IEEE Visualization 2005, Proceedings. 2005. 1566031 https://doi.org/10.1109/VIS.2005.73
Tzeng, Fan Yin ; Ma, Kwan-Liu. / Opening the black box -Data driven visualization of neural networks. VIS 05: IEEE Visualization 2005, Proceedings. 2005.
@inproceedings{0ddb0ddad9544915978be18989d8f6d1,
title = "Opening the black box -Data driven visualization of neural networks",
abstract = "Artificial neural networks are computer software or hardware models inspired by the structure and behavior of neurons in the human nervous system. As a powerful learning tool, increasingly neural networks have been adopted by many large-scale information processing applications but there is no a set of well defined criteria for choosing a neural network. The user mostly treats a neural network as a black box and cannot explain how learning from input data was done nor how performance can be consistently ensured. We have experimented with several information visualization designs aiming to open the black box to possibly uncover underlying dependencies between the input data and the output data of a neural network. In this paper, we present our designs and show that the visualizations not only help us design more efficient neural networks, but also assist us in the process of using neural networks for problem solving such as performing a classification task.",
keywords = "Artificial Neural Network, Classification, Information Visualization, Machine Learning, Visualization Application",
author = "Tzeng, {Fan Yin} and Kwan-Liu Ma",
year = "2005",
month = "12",
day = "1",
doi = "10.1109/VIS.2005.73",
language = "English (US)",
isbn = "0780394623",
booktitle = "VIS 05",

}

TY - GEN

T1 - Opening the black box -Data driven visualization of neural networks

AU - Tzeng, Fan Yin

AU - Ma, Kwan-Liu

PY - 2005/12/1

Y1 - 2005/12/1

N2 - Artificial neural networks are computer software or hardware models inspired by the structure and behavior of neurons in the human nervous system. As a powerful learning tool, increasingly neural networks have been adopted by many large-scale information processing applications but there is no a set of well defined criteria for choosing a neural network. The user mostly treats a neural network as a black box and cannot explain how learning from input data was done nor how performance can be consistently ensured. We have experimented with several information visualization designs aiming to open the black box to possibly uncover underlying dependencies between the input data and the output data of a neural network. In this paper, we present our designs and show that the visualizations not only help us design more efficient neural networks, but also assist us in the process of using neural networks for problem solving such as performing a classification task.

AB - Artificial neural networks are computer software or hardware models inspired by the structure and behavior of neurons in the human nervous system. As a powerful learning tool, increasingly neural networks have been adopted by many large-scale information processing applications but there is no a set of well defined criteria for choosing a neural network. The user mostly treats a neural network as a black box and cannot explain how learning from input data was done nor how performance can be consistently ensured. We have experimented with several information visualization designs aiming to open the black box to possibly uncover underlying dependencies between the input data and the output data of a neural network. In this paper, we present our designs and show that the visualizations not only help us design more efficient neural networks, but also assist us in the process of using neural networks for problem solving such as performing a classification task.

KW - Artificial Neural Network

KW - Classification

KW - Information Visualization

KW - Machine Learning

KW - Visualization Application

UR - http://www.scopus.com/inward/record.url?scp=33749429989&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33749429989&partnerID=8YFLogxK

U2 - 10.1109/VIS.2005.73

DO - 10.1109/VIS.2005.73

M3 - Conference contribution

SN - 0780394623

SN - 9780780394629

BT - VIS 05

ER -