Deep Learning in Practice

MVA Master, January-March 2022

News : (reload for fresher news...)
- Comments welcomed! about this class here

The new course (January-March 2023) can be found there.

General information about the course: presentation slides

Requirements : having already followed a course about neural networks (this is an advanced deep learning course).
Typical mathematical notions used: differential calculus, Bayesian statistics, analysis, information theory.

Teaching team :
In a nutshell:
Chapter Title Summary Notes Videos Exercises
1 Deep learning vs. classical ML and optimization html pdf 1 Hyperparameter and training basics
Environment and practical tips and tricks
2 Interpretability html pdf 1, 2, 3, 4, 5 Visualization with grad-CAM: notebook and test images
3 Architectures html pdf 1, 2 Graph-NN: instructions, code and baseline
4 Small data: weak supervision, transfer, and incorporation of priors html pdf 1, 2 Transfer learning: jupyter notebook
5 Deep learning and physics (invariances, priors or physical properties)
with Michele Alessandro Bucci (Safran)
html + refs pdf 1, 2 Learning dynamical systems
6 Generative models
+ uncertainty quantification by Maxence Ernoult (IBM Research)
- pdf 1, 2
Generative models
7 Modeling tasks and losses
+ Generalization and guarantees
+ Auto-ML / Auto-DeepLearning by Adrian El Baz, from MILA
html 1 + 2 pdf 1, 2

Schedule : January-March 2022, on (most) Thursday mornings, 9h - 12h15 (3 hours with a 15 minute break)
Location : now hybrid mode: online (same link as usual) + at CentraleSupelec in amphi Janet (also named room E2.19, in the Breguet building)
Location info: bus stop "Moulon" + 7 minute walk to the Breguet building. Common transportation planification: vianavigo.

Ressources: links to introduction to python, numpy, classical machine learning + online deep learning courses

Internship offers :

Back to the main page

Valid HTML 4.0 Transitional