Production RL and Decision-Making with RLlib

Learn about production reinforcement learning (RL) and decision-making using RLlib in this 29-minute conference talk from Anyscale's Ray Summit 2022. Explore real-world applications of RL in industries such as gaming and chemicals, and discover how to implement recommendation systems. Address common challenges in RL, including problem formulation, constraints, evaluation, and deployment. Gain insights into utilities, side-by-side code comparisons, and techniques for running evaluations and serving RL models. Understand the complexities of implementing RL solutions in production environments and get answers to your questions about practical RL applications.

Intro
Who am I
Who are we
Overview
Riot Games
Dow Chemicals
Recommendation systems
Challenges
The most basic challenge
Constraints
Evaluation
Deployment
Problem formulation
Utilities
Sidebyside code
Running evaluations
Serving RL models
Ideas are easy
Questions