Deep Learning Project - 4A
Resources for the 4th Year's Deep Learning Project at Polytech Lyon
Subjects:
Sujet 1 : Problème inverse - Tomographie par émission
Sujet 2 : Apprentissage par renforcement - équilibrage d'un double pendule
Sujet 3 : Classification et segmentation faiblement supervisée de tissus cancéreux
Some videos for theoretical background:
You can start by watching But what is a neural network REALLY? by Algorithmic Simplicity. To go further, the entire 3Blue1Brown neural networks playlist is a great starting point. Once you have finished that playlist, you can return to Algorithmic Simplicity's channel and look for videos about specific models.
The first two channels focus heavily on mathematics, so if you are looking for a perspective more closely aligned with computer science, consider watching Sebastian Lague's video: How to Create a Neural Network (and Train it to Identify Doodles).
Before jumping into PyTorch and TensorFlow, try building your own neural network from scratch to check that you fully understand the basics of backpropagation and neural networks.