Introduction to Probabilistic Graphical Models and Deep Generative Models

Pierre Latouche - Pierre-Alexandre Mattei
UCA+Ecole Polytechnique, INRIA

Master recherche specialité "Mathématiques Appliquées",
Parcours M2 Mathématiques, Vision et Apprentissage (ENS Paris-Saclay), 1er semestre, 2023/2024

Lectures will be given on Wednedays (9-12am). This year, the students will be evaluated by a [report + poster presentation] on a research paper. All lectures will be on zoom and recorded. The lectures at ENS Paris Saclay will be used to answer questions in person, regarding the lectures and the project. All documents (slides / blackboard / lecture notes)

Description

This course provides a unifying introduction to probabilistic modelling through the framework of graphical models, together with their associated learning and inference algorithms. It is one of the few historical courses at the core of the MVA program. Recent developments in deep latent variable models such as deep neural networks, variational autoencoders, generative adversarial networks are now part of the program of this course.

References

The course is based on the Pattern Recognition and Machine Learning book by C. Bishop . It is also based on the books Probabilistic Machine Learning: An Introduction and Probabilistic Machine Learning: Advanced Topics by K. Murphy.

Projects

List of research papers for the project.

Form to register a group and choose a research paper.

Internship offers

Internship: how to aggregate deep ensembles?

This internship will be supervised by Pierre-Alexandre Mattei (Research scientist at Inria, Maasai team), and will involve collaborations with Damien Garreau (Associate prof. at Université Côte d’Azur, also a member of Maasai). More details here.

Internship: learning why missing values are missing

To apply, please contact Pierre-Alexandre Mattei (pierre-alexandre.mattei\@inria.fr) and Aude Sportisse (aude.sportisse\@inria.fr). More details here.

Dates of classes

Date	Lecturer	Topics
October 4th zoom	Pierre Latouche	Introduction Maximum likelihood Linear regression Logistic regression
October 11th ENS Paris Saclay (salle 1Z18, bâtiment Nord 1er étage) + zoom	Pierre Latouche	K-means EM Gaussian mixtures PPCA
October 18th zoom	Pierre Latouche	Bayesian linear regression Gaussian processes EM revisited Model selection
October 25th ENS Paris Saclay (salle 0I10, bâtiment Ouest- RDC) + zoom	Pierre-Alexandre Mattei	Directed graphical models: theory and examples
November 8th zoom	Pierre-Alexandre Mattei	Undirected graphical models Sum-product algorithm HMM
November 15th zoom	Pierre Latouche	Approximate inference I: variational techniques Stochastic block models + VEM Expectation propagation
November 22nd zoom	Pierre-Alexandre Mattei	Approximate inference II: Monte Carlo, MCMC
December 6th ENS Paris Saclay (salle 0I10, bâtiment Ouest- RDC) + zoom	Pierre-Alexandre Mattei	Approximate inference III: amortized variational inference Deep latent variable models, variational auto-encoders
December 13th zoom	Pierre-Alexandre Mattei	Deep generative models beyond VAEs GAN, autoregressive models, normalizing flows, ...
January 10th ENS Paris Saclay (salle 0I10, bâtiment Ouest- RDC)		Poster session

Last updated: November 15th, 2023.