Probabilistic Generative Models

This is the website for the Probabilistic Generative Models course (TC4) of Master 2 AI (2020-2021).

Important notice

If you follow this class, please send me an email at caio.corro@u-psud.fr:

• Object: [TC4] Registration
• Content: your name, whether you speak french or not and whether you are able to attend to in-person classes

I will then register you to the mailing list of the course

About

Recently, generative models have (again) become a hot topic in machine learning thanks to recent advances in deep learning. One of the benefit of these models is their ability to generate new data, see for example:

• face generation: https://thispersondoesnotexist.com/
• word generation: https://thisworddoesnotexist.com/

Moreover, they can be used for semi-supervised learning, feature extraction via latent variables, …

In this course, we will first review the theoretical background required to understand modern generative models:

• Probabilities,
• Latent variables models,
• Expectation Maximization algorithm,
• Change of variable theorem,
• Neural parameterization of probability distributions.

Based on this, we will study modern generative models based on neural networks:

• Variational Auto-Encoders (VAEs),
• Generative Adversarial Networks (GANs),
• Flow models,
• Energy networks.

Grading Scheme

• 50%: Lab exercises
• 50%: Exam

Contact

You can contact me at caio.corro@u-psud.fr, either in French or English, with a subject starting with [TC4]. Please, do not worry about typos or not being overly formal enough (just treat your instructors and colleagues with the same respect you would like to be treated).

WARNING: Each mail must discuss at most one point. Don’t send e-mails to several of my addresses. Thank you.