Lectures

This page contains links to all the slides used during the lectures. Click on the title and they will open them directly in your browser.

• Lecture 01. Course logistics and the success to deep learning
  PDF   Week 1
  
  Description: We quickly describe the course logistics, motivate deep learning by illustrating several of its numerous successes in many fields and explain quickly some ingredients for this success.
• Lecture 02. Machine learning recap, history of neural networks and the main building blocks
  PDF   Week 1
  
  Description: We give a quick recap about machine learning, tell a bit about the history of neural networks and explain the main building blocks: the multi-layer perceptron, stochastic gradient descent, computational graphs and the backpropagation algorithm..
• Lecture 03: Some hyperparameters, regularization techniques and practical recommendations
  PDF   Week 2
  
  Description: This lecture is about activation functions, output units and losses, initialization of weights, regularization by penalization, Dropout, batch and layer normalization, early stopping and will give general practical recommendations to train a neural net
• Lecture 04: About optimization and optimizers
  PDF   Week 2
  
  Description: This lecture is focuses on optimizers used for training neural networks together with momentum and learning rate scheduling techniques
• Lecture 05: Convolutional neural networks or how to make a neural net see ?
  PDF   Week 3
  
  Description: We give a little history about convolutional neural networks, explain convolution layers, pooling, some convolutional neural nets architectures and explain how one can look under the hood of such architectures.
• Lecture 06: Computer vision: architectures for the main tasks in computer vision
  PDF   Week 3
  
  Description: We will explain quickly the classification, detection and segmentation tasks in computer vision and the main architecture used for them.
• Lecture 07: Recurrent neural networks
  PDF   Week 4
  
  Description: We will explain how to learn from sequence using neural networks, including architectures such as RNN, LSTM and GRU.
• Lecture 08: Transformers, attention models and self-supervised learning
  PDF   Week 5
  
  Description: We will finish this course with some advanced topics in deep learning's current state-of-the-art.