Practical Artificial Intelligence for A/B Testing [Video]

More Information
  • Deploy an AI agent to perform A/B test between two versions of a simple webpage.
  • Explore Reinforcement Learning topics such as agent, environment, actions, and rewards.
  • Discover how to solve Multi-Armed bandit problem and use it for A/B test.
  • Apply Reinforcement Learning in a real-world use case besides the traditional games examples.
  • Clarify once for all the difference between Exploration and Exploitation in an AI context.
  • Design, code and experiment different strategies for AI agent in Python.
  • Deploy the agent to perform a real A/B test using Flask, a web framework.

A/B testing is a well-known technique in web designing where designers apply it to test out different versions of the same webpage. The drawback to this technique is the waiting time to choose the best version and you lose the current performance of the webpage. To counter these drawbacks, you will learn how to build an AI Agent to A/B test the webpage in a much quicker pace using Reinforcement Learning.

This course will teach you how to build and deploy an AI Agent to test multiple versions of the web page and choose the best one much faster than the traditional A/B testing method. This quick decision-making will ensure good performance of your web-page even during the experiment.

By the end of this course, you will be able to deploy an AI Agent to perform an A/B test with many different strategies and to select the one which boosts its performance.

The code bundle for this video course is available at-

Style and Approach

This course is a full hands-on tutorial. It shares only the essential theoretical concepts that will context and support the practical implementation of the solution. The essential theory is explained using examples, analogies, and figures to make it self-explanatory and as clear as possible. In the practical portion, all the codes are developed on the fly during the videos, each line is explained with minimum jargon. The main object is to make the viewers succeed and overcome obstacles in their own environments.

  • Improve the outcome of your A/B test, using clear and simple Python implementations of the concepts of AI and Reinforcement Learning.
  • Learn how to try out different solutions to solve the Multi-Armed bandit problem which is one of the most import challenges of Reinforcement Learning.
  • The course is focused on the resolution of the challenge. Each step is composed of two portions. The first one is a brief and essential theoretical context and the second is the practical development of the solution. 
Course Length 3 hours 3 minutes
ISBN 9781788990745
Date Of Publication 28 Feb 2019


Meigarom Diego Fernandes Lopes

Meigarom Diego Fernandes Lopes is a Senior Data Scientist that has been working on data projects for 4 years. He is an expert in deploy Machine Learning models to solve business problems and support decision-makers with cut-edge technologies. His primary interests are in Reinforcement Learning, Recommendation Systems, Machine learning models in general and data engineering process. He has worked on game-changing projects like NLP algorithms to measure satisfaction of customer about the service through comments of the E-hailing company, recommendation system able to suggest financial products to invest money and he is currently working on an fashion e-commerce executing projects like fashion model image classification, product auto-tagging, and sales forecast. Finally, he writes about Machine learning models on his own Medium blog.