Get a clear understanding of current AI options
Get a clear understanding of the technologies involved in neural networks
Get a clear understanding of the neural networks as function estimators
Get a clear understanding of the data pre-processing issue
Ability to assess the best solution (between classical AI and ML) for a given problem
Ability to choose between supervised learning and reinforcement learning for a given problem
Ability to implement and tune a neural network option for a given problem
Ability to gather and prepare data for an AI implementation
Day 1: Starter Content
Day 2: Essential Content
Mixing Reinforcement Learning and Deep Learning
Estimating the Q table via neural networks
Application for Atari Pong game
Tuning a Convolutional Neural Network
Matrix convolution and filters
Main parameters: activation functions, pooling layers, drop out layers
Hyper-parameters tunings such as numbers of filter, stride and padding
Preparing data: Cleaning and Normalisation
Facebook Artificial Intelligence Research
Gilles Richard is a Professor of Computer Science at Paul Sabatier University (UPS – Toulouse 3 – France) where he is involved in undergraduate curricula and master programs. He is also the Founder of Geom, an AI based startup that focuses on transforming 2-dimensional images in 3-dimensional objects using machine learning algorithm and image processing.
He is also involved in Dystech, Dystech goal is to detect Dyspraxia by analysing a simple handwritten text, also using machine learning and image recognition.
Gilles is also the author of the book Computational Approaches to Analogical Reasoning.
He discusses e-commerce’s impact on the retail industry, key things customers expect from an e-commerce platform, main obstacles restricting the Middle East’s e-commerce industry, and more.