https://towardsdatascience.com/facebook-open-sources-horizon-to-streamline-the-implementation-of-reinforcement-learning-solutions-1d39414bd689

Reinforcement learning is one of the most exciting areas of development in the current artificial intelligence(AI) landscape. From AlphaGo to OpenAI Five, reinforcement learning has been at the center of major AI breakthroughs in the last few years. And yet, the implementation of reinforcement learning remains difficult enough that only specialized teams with advanced AI research skills have been able to pursue those efforts. Not surprisingly, most of the interesting reinforcement learning applications we see today come from AI powerhouses like Google, Microsoft, Amazon, Apple or Facebook. In the case of Facebook, the social media giant has been using reinforcement learning across different scenarios such as intelligent notifications or the M assistant. Recently, the Facebook engineering team open sourced Horizon, a framework that brings together some of the best practices Facebook has learned in the implementation of reinforcement learning solutions so that they can be used by mainstream developers.

The release of Horizon was accompanied by the publication of a research paper that describes the core principles of the framework. Horizon doesn’t only focus on simplifying the implementation of reinforcement learning solutions but specially targets large scale scenarios that involve factors such as large datasets with hundreds or thousands of varying feature types and distributions, high dimensional discrete and continuous action spaces and other aspects that make reinforcement learning nothing short of a nightmare. More concretely, the Facebook team specified six key goals for Horizon: