Are you an EPFL student looking for a semester project?
Work with us on data science and visualisation projects, and deploy your project as an app on top of Graph Search.
This lecture covers the fundamentals of deep learning, focusing on Convolutional Neural Networks (CNNs). It starts with a recap of neural network basics, then delves into the architecture and training of CNNs. The lecture explains the concept of transposed convolutions for fully-convolutional networks and explores standard CNN architectures like LeNet-5, AlexNet, VGG, and ResNet. It also discusses semantic segmentation using CNNs and tricks of the trade in deep learning, such as pre-training, learning rate scheduling, cyclic learning rates, and data augmentation. The lecture concludes with a discussion on adversarial examples in deep networks.