Global training of document processing systems using graph transformer networks

Leon Bottou, Yoshua Bengio, Yann LeCun

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

Abstract

We propose a new machine learning paradigm called Graph Transformer Networks that extends the applicability of gradient-based learning algorithms to systems composed of modules that take graphs as inputs and produce graphs as output. Training is performed by computing gradients of a global objective function with respect to all the parameters in the system using a kind of back-propagation procedure. A complete check reading system based on these concepts is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provide record accuracy on business and personal checks. It is presently deployed commercially and reads millions of checks per month.

Original languageEnglish (US)
Title of host publicationProceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition
Editors Anon
PublisherIEEE
Pages489-494
Number of pages6
StatePublished - 1997
EventProceedings of the 1997 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - San Juan, PR, USA
Duration: Jun 17 1997Jun 19 1997

Other

OtherProceedings of the 1997 IEEE Computer Society Conference on Computer Vision and Pattern Recognition
CitySan Juan, PR, USA
Period6/17/976/19/97

Fingerprint

Processing
Backpropagation
Learning algorithms
Learning systems
Neural networks
Industry

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Software
  • Control and Systems Engineering
  • Electrical and Electronic Engineering

Cite this

Bottou, L., Bengio, Y., & LeCun, Y. (1997). Global training of document processing systems using graph transformer networks. In Anon (Ed.), Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (pp. 489-494). IEEE.

Global training of document processing systems using graph transformer networks. / Bottou, Leon; Bengio, Yoshua; LeCun, Yann.

Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. ed. / Anon. IEEE, 1997. p. 489-494.

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

Bottou, L, Bengio, Y & LeCun, Y 1997, Global training of document processing systems using graph transformer networks. in Anon (ed.), Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE, pp. 489-494, Proceedings of the 1997 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, San Juan, PR, USA, 6/17/97.
Bottou L, Bengio Y, LeCun Y. Global training of document processing systems using graph transformer networks. In Anon, editor, Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. IEEE. 1997. p. 489-494
Bottou, Leon ; Bengio, Yoshua ; LeCun, Yann. / Global training of document processing systems using graph transformer networks. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition. editor / Anon. IEEE, 1997. pp. 489-494
@inproceedings{59ceda8e9ecd4a7dae8ab64f96e40dfc,
title = "Global training of document processing systems using graph transformer networks",
abstract = "We propose a new machine learning paradigm called Graph Transformer Networks that extends the applicability of gradient-based learning algorithms to systems composed of modules that take graphs as inputs and produce graphs as output. Training is performed by computing gradients of a global objective function with respect to all the parameters in the system using a kind of back-propagation procedure. A complete check reading system based on these concepts is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provide record accuracy on business and personal checks. It is presently deployed commercially and reads millions of checks per month.",
author = "Leon Bottou and Yoshua Bengio and Yann LeCun",
year = "1997",
language = "English (US)",
pages = "489--494",
editor = "Anon",
booktitle = "Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition",
publisher = "IEEE",

}

TY - GEN

T1 - Global training of document processing systems using graph transformer networks

AU - Bottou, Leon

AU - Bengio, Yoshua

AU - LeCun, Yann

PY - 1997

Y1 - 1997

N2 - We propose a new machine learning paradigm called Graph Transformer Networks that extends the applicability of gradient-based learning algorithms to systems composed of modules that take graphs as inputs and produce graphs as output. Training is performed by computing gradients of a global objective function with respect to all the parameters in the system using a kind of back-propagation procedure. A complete check reading system based on these concepts is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provide record accuracy on business and personal checks. It is presently deployed commercially and reads millions of checks per month.

AB - We propose a new machine learning paradigm called Graph Transformer Networks that extends the applicability of gradient-based learning algorithms to systems composed of modules that take graphs as inputs and produce graphs as output. Training is performed by computing gradients of a global objective function with respect to all the parameters in the system using a kind of back-propagation procedure. A complete check reading system based on these concepts is described. The system uses convolutional neural network character recognizers, combined with global training techniques to provide record accuracy on business and personal checks. It is presently deployed commercially and reads millions of checks per month.

UR - http://www.scopus.com/inward/record.url?scp=0030648914&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0030648914&partnerID=8YFLogxK

M3 - Conference contribution

SP - 489

EP - 494

BT - Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition

A2 - Anon, null

PB - IEEE

ER -