Environment

The accelerated decision making with AI tool has implemented the environments in a OpenAI gym styled environment.

The accelerated decision making with AI environment is novel in that it features a client-server architecture which supports sharing of resource intensive elements, as well as providing a mechanism to support transparency and trust.

To make the problem tractable and repeatable, we capture output from the models into a form which preserves the stochasticity and intervention dimensionality of the domain. This forms the basis for an environment which can be used to benchmark the performance of various learning algorithms.

The environment is an object which provides a surrogate for the complex domain model, and is a basis for learning the utility of performing given interventions at specific times.

This formulation is one which is well-suited for the application of Artificial Intelligence (AI) agents to learn the most effective intervention strategies for a specific environment.

Different environments can be used to callibrate or intervention planning for a single model by altering the input sample space or/and output/observation sample space.