1.8. AI Engine Overview

1.8.1. Overview

The Citrine Platform uses artificial intelligence (AI) to accelerate your materials research and development. The platform exposes several types of Modules – re-usable assets that codify domain knowledge, research capabilities, or experimental objectives. Modules are combined into Workflows that provide actionable information to help direct research and development.

The Citrine Python client allows you to create and manipulate Modules and Workflows. Workflows can be executed and the results inspected. This page provides the briefest of overviews; for a more thorough discussion see the AI Engine documentation.

1.8.2. Modules

1.8.2.1. Predictors

Predictors define relations between variables. Predictors can be machine learning models, analytic relations, or featurizers. Several Predictors can be combined into a graphical model that expresses your domain knowledge and predicts material properties with quantified uncertainty estimates.

1.8.2.2. Design Spaces

Design Spaces define the set of materials of interest.

1.8.3. Workflows

1.8.3.1. Design Workflow

Design Workflows generate proposals for materials that are expected to meet some goal. A Design Workflow combines a Design Space to define the materials of interest and a Predictor to predict material properties. They also include a Score which codifies goals of the project.

1.8.3.2. Predictor Evaluation Workflow

Predictor Evaluation Workflows analyze the quality of a Predictor.