This course follows the Machine Learning and the High Dimensional & Deep Learning. In theses courses, you have acquired knowledge in machine and deep learning algoritms and their application on various type of data. This knowledge is primordial to become a DataScientist.

As a DataScientist, you will also need to know the tool that wil allow you to perform these algorithms efficiently. The two main goals of this course are :

• Discover these different tools:
  • Distributed computation with Spark.
  • Cloud computing with Google Cloud.
  • Container with Docker.
• Use this tools on various domain of learning with real dataset and with usefull learning librairy.
  • Natural language processing with Nltk, Scikit-Learn and Gensim
  • Recommendation system.
  • Reinforcement Learning with Gym Open AI

You will follow introduction to these different technologies.

• R Tutorial
• Python Tutorial
• Elementary statistic tools
• Data Exploration and Clustering.
• Machine Learning
• High Dimensional & Deep Learning

The course is divided in 5 sessions.

• Session 1 - 04-11-19 Spark
  • Python complement
  • Introduction to Spark via API PySpark API.
• Session 2 - 25-11-19 Cloud computing and containerization.
  • Configure and start an Google Cloud instance.
  • Build Docker Image and run container.
• Session 3 - 02-12-19 NLP (Natural Language Processing)
  • Cdiscount dataset : Classification of product description
  • Text cleaning, Vectorization, Words Embedding, Supervised classification, RNN.
• Session 4 - 16-12-19 Reinforcment Learning
  • Use open AI environment
  • Policy gradient, Q-Learning.
• Session 5 Recomentdation System
  • MovieLens dataset.

You will be evaluated on your capacity of acting like a Data Scientist, i.e.

• Choose an algorithm you haven't seen during course understand it.
• Make it run on an dataset to evaluate its performances.
• Make it run on the appropriate tools (SPark? Cloud? GPu?)
• Share it and make your results easily reproducible (Git - docker? , conda environment?).

Examen

1. Project - (50%): A Git repository where
   • All code will be available to easily reproduce your result
   • Instruction will be clear
   • Deadline : January 11th, 2019.
2. Oral presentation - (50%):
   • In-Deep explanation of the chosen algorithm.
   • Comment results (with critical mind)
   • Choice of the tools-infrastructure used.
   • Difficulty you've met.
   • Date : January 14th, 2019.