Program

(please click here for the videos and slides of the oral presentations.  Login and password was sent to you by e-mail. If not, please contact pierre-marc.jodoin [at] usherbrooke.ca or thomas.grenier [at] insa-lyon.fr)

Monday July 4

8h30 - 9h30       Registration and Breakfast

9h30 - 9h45       Welcome talk

9h45 - 10h30     Introduction to machine learning Part 1 (Pierre-Marc Jodoin)
Basics of machine learning, classification vs regression, train and test sets, metrics, over and under fitting, etc.

10h30 - 10h45   Coffee Break

10h45 - 12h00     Introduction to machine learning Part 2 (Pierre-Marc Jodoin)
Basics of machine learning, classification vs regression, train and test sets, metrics, over and under fitting, etc.

12h00 - 13h30    Lunch at the ETS

13h30 - 15h00    Basics in deep learning 1 (Pierre-Marc Jodoin)
Perceptron and multi-layer perceptron, stochastic gradient descent, learning rate, logistic regression, activation function, regularization (L1/L2/dropout/early stopping), etc.

15h00 - 15h30     Coffee break

15h30 - 17h00     Basics in deep learning 2 (Christian Desrosiers)
Weights initialization, forward and backward propagation, batch size, convolution neural nets (CNN), feature maps, pooling, pretraining and transfer learning, applications.

17h30 - 19h30     Cocktail at the ETS

Tuesday July 5

8h30 - 9h00     Breakfast

9h00 - 10h30   Advanced concepts in deep learning 1 (Michaël Sdika)
Common CNN architectures for classification (VGGNet, ResNet, ...) and localization (FasterRCNN, Yolo) and segmentation (encoder-Decoder, U-Net, ENet, ...)

10h30 - 11h00   Coffee Break

11h00 - 12h30   Generative and adversarial methods for medical imaging (Mohammad Havaei)
GANs, autoencoders and their training

12h30 - 14h00 Lunch at the ETS

15h00 - 15h30 Coffee break

14h00 - 17h00   Hands-on session 1: Introduction (Michaël Sdika, Thomas Grenier, Arash Ash, David Osowiechi)
Classification from machine learning to deep learning

Wednesday July 6

8h30 - 9h00     Breakfast

9h00 - 10h30    Advanced concepts in deep learning 2 (Hassan Rivaz)
Explainability, RNN, LSTM, Transformers, Self-supervised learning, AI-powered ultrasound, etc.

10h30 - 11h00   Coffee Break

11h00 - 12h30   Typical medical imaging issues (Samuel Kadoury)
Domain adaptation, privacy protection and federated learning, adversarial learning, common pitfalls, incomplete data, etc.

12h30 - 14h00   Lunch at the ETS

15h00 - 15h30    Coffee break

14h00 - 17h00   Hands-on session 2: Segmentation using deep learning  (Michaël Sdika, Thomas Grenier, Arash Ash, Mélanie Gaillochet)

18h00 - 21h30    Museum visit and banquet dinner

Thursday July 7

8h30 - 9h00       Breakfast

9h00 - 10h30     Is my model interpretable, explainable, valid and useful? (Ryeyan Taseen)

10h30 - 11h00   Coffee Break

11h00 - 12h30   Round table (Ryeyan Taseen, Jean-René Bélanger, Laurent Létourneau-Guillon, Mohammad Havaei)
Why so much AI in research, why still so few AI in clinic?

12h30 - 14h00   Lunch at the ETS

15h00 - 15h30   Coffee break

14h00 - 17h00   Hands-on session 3: Variational Autoencoder (Pierre-Marc Jodoin, Gustavo Vargas, Mélanie Gaillochet, Shambhavi Mishra)
Auto-encoders, convolutional auto-encoders, variational auto-encoders, latent spaces

Friday July 8

8h30 - 9h00    Breakfast

9h00 - 10h30  Weakly supervised deep learning (Jose Dolz and Ismail Ben Ayed)
Weakly supervised segmentation, constrained CNN losses, semantic segmentation, semi-supervised learning

10h30 - 11h00  Coffee Break

11h00 - 12h30  Geometric deep learning (Hervé Lombaert)
Spectral coordinates and representation, spectral deep learning, brain surface matching and parcellation

12h30 - 12h45  Closing remarks

12h45 - 14h00  Lunch at the ETS

15h00 - 15h30  Coffee break

14h00 - 17h00  Hands-on session 4: Weakly supervised learning (Pierre-Marc Jodoin, Gustavo Vargas, Arash Ash, Christian Desrosiers and Jose Dolz)