nov 1, 1998 - Gradient-Based Learning Applied to Document Recognition

Description:

LeCun et al. introduce LeNet-5, a pioneering convolutional neural network (CNN) architecture that demonstrated the power of deep learning for image classification. Applied to the MNIST handwritten digit recognition task, their approach outperforms traditional computer vision pipelines by learning directly from raw pixel data without manual feature engineering. The network used a combination of convolutional, subsampling, and fully connected layers, trained end-to-end using backpropagation. This work proved that gradient-based learning with deep architectures could scale effectively and deliver superior results, laying foundational groundwork for the modern deep learning revolution in computer vision.

Added to timeline:

Date:

Images: