Yann LeCun1, Bernhard E. Boser1, John S. Denker1, D. Henderson1, Richard Howard1, W. Hubbard1, L. D. Jackel1
1AT&T Bell Laboratories, Holmdel, NJ 07733 USA
Tóm tắt
The ability of learning networks to generalize can be greatly enhanced by providing constraints from the task domain. This paper demonstrates how such constraints can be integrated into a backpropagation network through the architecture of the network. This approach has been successfully applied to the recognition of handwritten zip code digits provided by the U.S. Postal Service. A single network learns the entire recognition operation, going from the normalized image of the character to the final classification.
