Slides

ImageNet Classification with Deep Convolutional Neural Networks

Neural networks

Deep learning

Image classification

Convolutional neural networks

2012

This slide summarizes the landmark 2012 paper by Alex Krizhevsky, Ilya Sutskever, and Geoffrey E. Hinton. This work introduced AlexNet, a deep convolutional neural network that dramatically improved image classification performance on the ImageNet Large Scale Visual Recognition Challenge (ILSVRC). Leveraging GPU acceleration, ReLU activations, data augmentation, and dropout for regularization, AlexNet achieved a top-5 error rate of 15.3%, vastly outperforming previous methods. This breakthrough validated the power of deep learning at scale and marked the beginning of the modern era of deep convolutional neural networks in computer vision.

Gradient-based learning applied to document recognition

Neural networks

Character recognition

Convolutional neural networks

1998

This slide summarizes the landmark 1998 paper by Yann LeCun and colleagues, which introduced the LeNet-5 convolutional neural network architecture and demonstrated its effectiveness for handwritten digit recognition. The paper established convolutional neural networks (CNNs) as powerful models for image-based tasks, highlighting their ability to learn hierarchical features directly from raw pixels using gradient-based training. By applying CNNs to document recognition, the authors set a new benchmark on the MNIST dataset and laid the groundwork for the modern deep learning revolution.