Deep Reinforcement Learning with AnyLogic and Microsoft Project Bonsai

Project Bonsai End-to-End Workflow webinar

Microsoft Project Bonsai enables subject matter experts, even those with no AI background, to utilize their expertise to teach AI agents and solve real-world business problems. Here are our webinar recording and accompanying resources that introduce the key concepts and workflow for using Project Bonsai and AnyLogic simulation together.

The webinar follows AnyLogic and Microsoft joining forces to bring deep reinforcement learning and machine teaching to business applications. David Coe, principal program manager in the Microsoft Project Bonsai team and AnyLogic AI Program Lead, Arash Mahdavi, demonstrate the end-to-end workflow for training AI using simulation and for implementing AI control in a simulation.

AnyLogic and Project Bonsai End-to-End Workflow

The AnyLogic and Project Bonsai End-to-End Workflow webinar from Arash Mahdavi and David Coe.

During the webinar, to demonstrate training an AI agent in simulation, a refactored and wrapped Activity Based Costing Analysis example model is connected to the Bonsai platform. You can get a ready-to-go version of the model from our Project Bonsai page, where you will also find a second example model and the wrapper for using with your own simulation models. Included with the wrapper is a user guide document. [If you need access to Project Bonsai, apply for the Project Bonsai Preview]

To abstract away much of the complexity of deep reinforcement learning, Project Bonsai specifies training using Goals, Concepts, and Curricula. Together, these form the machine teaching instructions for training a Project Bonsai Brain and are specified declaratively using Inkling. David introduces the language before showing how to create a Project Bonsai Brain and start its local or cloud-based training.

After training, a ‘brain’ can control a simulation model. The webinar demonstrates this by returning to the Activity Based Costing Analysis simulation model in AnyLogic and controlling it with a brain that has been exported as an Azure web app.

If you have any questions regarding the webinar, please ask in the comments below or on the video’s YouTube page. You can also get in touch with us directly if you have any questions in general about AI and AnyLogic, just head over to our contact page. If you’d like to learn more, get our AI & Simulation in Business White Paper pack.

Keep up to date with AnyLogic webinars, courses, and events on our Events and Training Calendar.

Related posts