Lectures in course PHYS-467: Machine learning for physicists

Covers the back-propagation algorithm for deep neural networks and the importance of locality in CNN.

Covers the implementation of fully connected neural networks on two datasets using PyTorch.

Covers the motivation and architecture of convolutional networks, from LeNet to AlexNet.

Explores the motivation and ideas behind convolutional networks, emphasizing weight sharing and pooling layers.

Covers PyTorch tensor data structure and training a CNN to classify images.

Explores the visualization of feature maps in neural networks and the features learned by deep models.