Monday, 
January 16th

Tuesday, 
January 17th

Wednesday, January18th

9:00 - 12:45 

1.Introduction to machine learning

3. Introduction to time series forecasting and segmentation

4. Introduction to Natural Language Processing 

12:45 - 13:30

Lunch

13:30-17:00

See details for day

2. Introduction to deep learning in the industry

3. Introduction to time series forecasting and segmentation

5. Reinforcement Learning in industry

Session 1 -  Monday, January 16th
Theory: Alex Nguyen – from 9 AM to 11 AM

• Types of learning: supervised, unsupervised
• Introduction: Regression and Classification: Machine
  learning models vs. statistical approaches
• Introduction: Some examples and models
• Good practices: Overfitting and regularization
• Good practices: Experimental design

Tutorial: Rosnel Sessinou – [Solving insurance problems with ML]
from 11:15 AM to 12:45 PM

• Exploratory data analysis
• Regression models and Classification
• Generalization and regularization

Learning Objectives
At the end of these sessions, the participant will have reviewed:

• The basic concepts and good practices underlying machine learning
• The basic libraries that are most used when applying ML in the industry  

Session 2-  Monday, January 16th
Theory: Alex Nguyen – from 1:30 PM to 3:30 PM

• Multilayered Perceptron
• Introduction to Deep Neural Network Architectures (CNN, AE)

Tutorial: Pierre Rosin – [Inventory count - car in a parking lot] from 3:45 AM to 5:15 PM

• Deep / Representation Learning applied in image classification (ex., inventory count)
• How representation learning could allow boosting models
• How to correctly apply neural network in the context of image classification

Learning Objectives
At the end of these sessions, the participant will have reviewed:

• The basic concepts and good practices underlying deep learning
• The basic libraries that are most commonly used when applying deep learning in industry
• To demystify the use of deep learning in industry and the context required for deep learning applications to have an added value
• Specific use cases where deep learning can be applied: examples from Supply chain, Logistics, and Finance

Session 3-  Tuesday January 17th
Theory: Alex Nguyen – from 9 AM to 11 AM

• Introduction to recurrent neural networks (continued)
  (RNN, LSTM, GRU)
• Forecasting problem

Tutorial: Rosnel Sessinou– [Demand forecasting for storage management] from 11:15 AM to 12:45 PM

• LSTM, GRU for times series prediction
• RNN for times series prediction

Session 4-  Tuesday January 17th
Theory: Pierre Rosin – [Causal ML Part I] -
from 1:30 PM to 3:PM

• EconML: Customer segmentation
• Estimation of heterogeneous treatment effect using Orthogonal machine learning
• Making policy decisions

Tutorial: Pierre Rosin – [Causal ML Part II] - from 3:15 PM to 4:45 PM

• EconML: A/B testing
• Estimation of heterogeneous treatment effect using instrumental variables and orthogonal machine learning
• Making policy decisions

Learning Objectives
At the end of these sessions, the participant will have reviewed:

• Time series forecasting: The basics concepts and good practices when applying deep learning in time series data
• Causal ML
• The estimation of Heterogeneous Treatment Effects

Session 5- Wednesday, January 18th
Theory: Alex Nguyen – from 9 AM to 11 AM

• Introduction to classical and modern NLP (RNN, LSTM, GRU)
• Text processing
• Language representations

Tutorial: Alex Nguyen – [Modeling sentiments for decision making]
from 11:15 AM to 12:45 PM

• Sentiment analysis
• Application of convolutional neural networks (CNN)

Learning Objectives
At the end of these sessions, the participant will have reviewed:

• Introduction to document processing
• Application of deep learning methods to textual data

Session 6- Wednesday, January 18th
Theory: Rosnel Sessinou - from 1:30 PM to 3:30 PM

• Introduction: learning by reinforcement
• What is the ‘reinforcement learning problem’?
• One popular reinforcement learning approach: Q-learning
• Reinforcement learning for large-scale problems
• Pitfalls of tabular methods of reinforcement learning
• Function approximators for value and action-value functions
• Introduction to deep Q-networks
• Real-life applications of deep reinforcement learning

Tutorial: Pierre Rosin – [Inventory Management with Q-learning] - from 3:45 PM to 5:15 PM

• Setting up an environment for reinforcement learning
• Application of Q-learning to learn a policy

Learning Objectives
At the end of these sessions, the participant will have reviewed:

• The basics concepts and good practices underlying reinforcement learning
• Some use cases where RL can be applied in path optimization in transport