What If AI Drove Your Operations?
Pathmind simplifies every step of adding AI to simulation models, from finding the best decisions in early training stages to validating those decisions and deploying them into real-world systems.
Build. Train. Deploy.
Pathmind is a SaaS application that leverages the cloud. Pathmind’s AI platform contains three components: Pathmind Helper, Pathmind Training, and Pathmind Serving.
Build AI with Pathmind Helper
Pathmind Helper features accessed through an easy-to-use properties panel in AnyLogic
Pathmind Helper helps you define and build AI in your simulations. Its simple user interface guides you through the steps to prepare simulation models to succeed with AI, including:
- Defining observations needed for AI to get information from a model.
- Setting up actions and specifying when those actions are triggered, allowing AI to interact with a model.
- Adding conditional or time-based criteria for when AI training in the simulation should end.
- Establishing reward variable building blocks that determine what the AI should learn or optimize.
Pathmind Helper also allows you to use trained decision-making policies within a model for simulation optimization and AI validation.
Extensive help documentation and tutorials are available for Pathmind Helper and no high-level AI knowledge or PhD are needed to use it. Pathmind Helper can be downloaded from the Resources page, and is also included with a Pathmind Training subscription.
Note: Pathmind Helper is available for AnyLogic and will be available for other simulation platforms and programming language environments soon.
Train AI with Pathmind Training
Manage projects and experiments with ease via the Pathmind Training interface
Pathmind Training is a web application that helps you leverage the power of cloud compute to train the AI created with Pathmind Helper in your simulation tool.
Many other AI platforms require special languages, have arcane infrastructure management technologies, and feature arcane command line APIs. Pathmind Training avoids those issues and lets you focus on getting results fast.
Pathmind Training’s user-friendly browser interface lets you:
- Upload models that have been set up for AI with Pathmind Helper to the cloud.
- Define experiments and test AI training criteria.
- Run multiple experiments simultaneously with cloud compute resources.
- Benefit from automatic selection of the best AI learning hyperparameters.
- View training results in a simple graph interface.
- Receive notifications when training completes.
- Export trained AI policies for validation, optimization, and real-world deployment with Pathmind Serving.
Deploy AI with Pathmind Serving
A business process logic engine calling Pathmind Serving REST API
Once a simulation has been optimized with AI and good results have been found, those decisions need to applied to real-world operations. Pathmind Serving is the easiest, fastest way to deploy trained AI into production. It enables teams to avoid the traditional pains associated with deploying AI and instead enables users to focus on business results by:
- Publishing trained AI policies as easy-to-secure web service in minutes.
- Using industry-best practice RESTful APIs familiar to modern developers.
- Providing REST resources for prediction, documentation, and testing.
- Integrating with business rules engines, logistics, and supply chain management system using industry best practices.
For app developers and IT teams working with legacy platforms or custom applications, Pathmind Serving also generates API documentation and example code clients in multiple programming languages, reducing the time for legacy and custom integrations to a matter of minutes.
Pathmind Serving takes the pain away from putting AI into production and lets you focus on putting the power of AI to work.
Note: Pathmind Serving is a separate subscription license from Pathmind Training.