Program
Introduction to ML/CV and basic topics in statistical learning - ERM, overfitting, kNN, cross-validation, classification/regression, error decomposition
Linear predictors - linear classifier, linear regression, logistic regression
Basic optimization - loss surface, computing the gradient, gradient descent, SGD, backpropagation
Neural networks basics - basic layers and activations, loss functions, regularizers
Advanced optimization and training - learning rate, momentum, Nesterov, ADAM, optimizing hyperparameters, train/validation graphs, intuitions and tips of training
Convolutional neural networks - convolutional layer, channels, stride, padding, advanced conv. layers, pooling layers; popular CNN architectures
Visualizing CNNs, adversarial examples - visualizing first layers, maximum activation of neurons, saliency maps, deep dream, adversarial examples, attacks and defenses
Advanced module - either generative models or RNNs
