Deep learning

This course focuses on deep learning approaches.

Introduction to supervised learning
Parametric approaches
Neural networks
"Perceptron" multilayer
Neural network parameter learning

Cost functions
Neural network parameter optimization by gradient backpropagation
Stochastic gradient descent

Parameter initialization parameters
Learning step definition
Learning step evolution
"Momentum"
ADAM




Premature termination
Neural network architecture

Convolution layer convolution
BatchNorm
Residual connection
ResNet


Neural network specialization
Data augmentation
"Adversarial examples"
Introduction to PyTorch
Application example : object detection