Abstract: 

Reinforcement Learning (RL) studies how an agent can interact with its environment to learn a policy that maximizes expected cumulative rewards for a task. Recently, RL has experienced growth and interest due to use case results in robotics, finance, autonomous driving, business management, education, energy, healthcare, to name a few. In this tutorial, we focus on a technique called Deep Q Learning. DQL is a RL technique where rewards are cataloged and maximized while learning the best policy. We provide a gentle introduction to Deep Q Learning with a canonical example of winning a game called the cart-pole problem. In the next part of the tutorial, we will outline how DQL applies to a finance example, predicting the next best move (i.e., best purchase and sales of stocks) to maximize gains. This valuable tutorial will teach both deep learning and distributed computing (Ray) skills.

Session Outline
Lesson 1: Deep Q Learning - what is it and how does it work?
Familiarize yourself with the general concepts of reinforcement learning, deep learning and q-learning. You will learn how these concepts tie together in the industrial context, with an emphasis on practical application.

Lesson 2: Learning about Ray and RLLib Libraries for Distributed RL
Practice using the distributed computing library, Ray, and its scalable reinforcement learning library, RLLib. We will have an initial tutorial where we solve the cart-pole problem and while doing so learn about distributed RL.

Lesson 3: Reinforcement Learning for Financial Predictions
We will discuss practical applications of reinforcement learning, and its use in financial institutions. We will walk through an example of applying deep q learning to making the best next choice for stock market purchases or sales. We will demonstrate how scalable techniques can make reinforcement learning a practical tool for business problems.

Background Knowledge
Attendees should have familiarity with Python and typical ML packages (e.g., pandas, numpy, sklearn, torch). We also suggest that attendees understand the basics of distributed computing, deep learning, and reinforcement learning. We will provide optional readings on the necessary background ahead of the tutorial.

Bio: 

Andrea Lowe, PhD is the Training and Enablement Engineer at Domino Data Labs where she develops training on topics including overviews of coding in Python, machine learning, Kubernetes, and AWS. She trained over 1000 data scientists and analysts in the last year. She has previously taught courses including Numerical Methods and Data Analytics & Visualization at the University of South Florida and UC Berkeley Extension. Her conference experience includes a deep learning tutorial at PyCon, 2 invited talks, 21 poster presentations, and 4 chair positions.