Deep reinforcement learning has been shown to outperform other optimization techniques on common operations research problems.
Pathmind provides an interface to implement and quickly train deep reinforcement learning agents within Python-based simulations.
This may be useful for people who:
Prefer to simulate real-life scenarios using Python code versus a simulation IDE that relies on a GUI and code snippets.
Prefer not to build and maintain their own deep reinforcement learning infrastructure and distributed training clusters.
Want to leverage Pathmind's automated hyperparameter tuning and algorithm selection.
To begin, you will need two things:
A use case conducive to reinforcement learning. For example, a use case that is stochastic in nature, in which actions can be chosen that impact an environment to influence outcomes.
A simulation built in Python.
Please contact Pathmind support to get started. Our Python implementation is currently in private beta.