Strong-RL is a framework for developing and deploying reinforcement-learning based action recommendation platforms at scale.
At its core, Strong-RL makes developing such applications easier in two ways:
By providing a conceptual framework with which to understand and approach your action recommendation problem.
By providing a set of Spark-based pipeline components that can easily be imported and extended into a new Python application to yield a highly-configurable, robust, and transparent solution.
With these tools, developing a reinforcement-learning based action recommendation engine that operates at web scale can be done in a single Python file. Or, they can form the basis of a larger application that uses Strong-RL as a foundation upon which to build a more complex solution.
In this documentation, we review some fundamental concepts about Strong-RL along with documentation of its API and internals.