Multilayer neural networks trained with the back-propagation algorithm
constitute the best example of a successful gradient based learning
technique. Given an appropriate network architecture, gradient-based
learning algorithms can be used to synthesize a complex decision
surface that can classify high-dimensional patterns, such as handwritten
characters, with minimal preprocessing. This paper reviews various
methods applied to handwritten character recognition and compares
them on a standard handwritten digit recognition task. Convolutional
neural networks, which are specifically designed to deal with the
variability of 2D shapes, are shown to outperform all other techniques.
Real-life document recognition systems are composed of multiple modules
including field extraction, segmentation recognition, and language
modeling. A new learning paradigm, called graph transformer networks
(GTN), allows such multimodule systems to be trained globally using
gradient-based methods so as to minimize an overall performance measure.
Two systems for online handwriting recognition are described. Experiments
demonstrate the advantage of global training, and the flexibility
of graph transformer networks. A graph transformer network for reading
a bank cheque is also described. It uses convolutional neural network
character recognizers combined with global training techniques to
provide record accuracy on business and personal cheques. It is deployed
commercially and reads several million cheques per day
%0 Journal Article
%1 Lecun1998
%A Lecun, Y.
%A Bottou, L.
%A Bengio, Y.
%A Haffner, P.
%D 1998
%J Proceedings of the IEEE
%K 2D GTN, back-propagation, backpropagation, based character cheque complex convolution, convolutional decision digit document extraction, field gradient gradient-based graph handwritten high-dimensional language learning learning, measure minimization, modeling, multilayer multimodule network networks, neural optical patterns, perceptrons, performance reading, recognition recognition, recognizers, segmentation shape surface synthesis, systems, task, technique, transformer variability,
%N 11
%P 2278--2324
%T Gradient-based learning applied to document recognition
%V 86
%X Multilayer neural networks trained with the back-propagation algorithm
constitute the best example of a successful gradient based learning
technique. Given an appropriate network architecture, gradient-based
learning algorithms can be used to synthesize a complex decision
surface that can classify high-dimensional patterns, such as handwritten
characters, with minimal preprocessing. This paper reviews various
methods applied to handwritten character recognition and compares
them on a standard handwritten digit recognition task. Convolutional
neural networks, which are specifically designed to deal with the
variability of 2D shapes, are shown to outperform all other techniques.
Real-life document recognition systems are composed of multiple modules
including field extraction, segmentation recognition, and language
modeling. A new learning paradigm, called graph transformer networks
(GTN), allows such multimodule systems to be trained globally using
gradient-based methods so as to minimize an overall performance measure.
Two systems for online handwriting recognition are described. Experiments
demonstrate the advantage of global training, and the flexibility
of graph transformer networks. A graph transformer network for reading
a bank cheque is also described. It uses convolutional neural network
character recognizers combined with global training techniques to
provide record accuracy on business and personal cheques. It is deployed
commercially and reads several million cheques per day
@article{Lecun1998,
abstract = {Multilayer neural networks trained with the back-propagation algorithm
constitute the best example of a successful gradient based learning
technique. Given an appropriate network architecture, gradient-based
learning algorithms can be used to synthesize a complex decision
surface that can classify high-dimensional patterns, such as handwritten
characters, with minimal preprocessing. This paper reviews various
methods applied to handwritten character recognition and compares
them on a standard handwritten digit recognition task. Convolutional
neural networks, which are specifically designed to deal with the
variability of 2D shapes, are shown to outperform all other techniques.
Real-life document recognition systems are composed of multiple modules
including field extraction, segmentation recognition, and language
modeling. A new learning paradigm, called graph transformer networks
(GTN), allows such multimodule systems to be trained globally using
gradient-based methods so as to minimize an overall performance measure.
Two systems for online handwriting recognition are described. Experiments
demonstrate the advantage of global training, and the flexibility
of graph transformer networks. A graph transformer network for reading
a bank cheque is also described. It uses convolutional neural network
character recognizers combined with global training techniques to
provide record accuracy on business and personal cheques. It is deployed
commercially and reads several million cheques per day},
added-at = {2009-09-12T19:19:34.000+0200},
author = {Lecun, Y. and Bottou, L. and Bengio, Y. and Haffner, P.},
biburl = {https://www.bibsonomy.org/bibtex/24d0e761be0edc4f631b3af679c2b33ab/mozaher},
file = {00726791.pdf:Lecun1998.pdf:PDF},
interhash = {7a82cccacd23cf06b25ff5325a6c86c7},
intrahash = {4d0e761be0edc4f631b3af679c2b33ab},
issn = {0018-9219},
journal = {Proceedings of the IEEE},
keywords = {2D GTN, back-propagation, backpropagation, based character cheque complex convolution, convolutional decision digit document extraction, field gradient gradient-based graph handwritten high-dimensional language learning learning, measure minimization, modeling, multilayer multimodule network networks, neural optical patterns, perceptrons, performance reading, recognition recognition, recognizers, segmentation shape surface synthesis, systems, task, technique, transformer variability,},
number = 11,
owner = {mozaher},
pages = {2278--2324},
timestamp = {2009-09-12T19:19:40.000+0200},
title = {Gradient-based learning applied to document recognition},
volume = 86,
year = 1998
}