Gradient-Based Learning Applied to Document Recognition

Yann LeCun, Leon Bottou, Yoshua Bengio, and Patrick Haffner published Gradient-Based Learning Applied to Document Recognition in Proceedings of the IEEE. The paper reviews gradient-based learning for document recognition, describes convolutional neural networks for two-dimensional shape variability, presents graph transformer networks for globally trained multimodule recognition systems, and documents the LeNet-5 architecture in its experimental treatment.