Deep Learning in Practice

MVA Master, January-March 2021

News : (reload for fresher news...)
- Link to hand in TD6 added (same password as previously)

The previous course (January-March 2020) can be found there; the new course will follow more or less the same lines, though not exactly.
The webpage for the new course (2021) is not completely finished; materials from last year and not updated yet are preceded by "[2020]".
General information about the course: presentation slides

Teaching team :
In a nutshell:
Chapter Title Summary Notes Videos Exercises
1 Deep learning vs. classical ML and optimization html pdf 1, 2 Hyperparameter and training basics
2 Interpretability html pdf 1, 2, 3 Visualization with grad-CAM
3 Architectures html pdf 1 Graph-NN: code and instructions
4 Small data: weak supervision, transfer, and incorporation of priors 1 + 2 pdf 1 Transfer learning: instructions and jupyter notebook
* Presentation of Therapixel, start-up in medical imaging, by Yaroslav Nikulin
Evolutionary Deep Computer Vision, by Olivier Teytaud (Facebook FAIR Paris)
slides + videos: 1, 2
5 Deep learning for physics, with Lionel Mathelin (LIMSI, Paris-Sud) and Michele Alessandro Bucci (LRI, TAU team) (html) slides 1, 2 Learning dynamical systems, with presentation
6 Modeling tasks and losses
+ Generative models (GAN, VAE and Normalizing Flows)
pdf 1
2, TP+3

Generative models
7 Guarantees? Generalization and formal proofs
+ Auto-ML / Auto-DeepLearning, by Lisheng Sun (Isabelle Guyon's group, LRI, Paris-Sud)

Schedule : January-March 2021, most often on Tuesday mornings, 9h - 12h15 (3 hours with a 15 minute break), online (note that the schedule is irregular): :

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