Genius documentation and tutorials (Darwin release)

What is Genius?

Genius is a Bayesian inference reasoning engine powered by active inference technology. With Genius you can build and query probabilistic models using your own domain knowledge or learn from a dataset. With active inference you can build a model that can take perform decision-making to choose the optimal course of action.

This is the documentation for the Darwin release of Genius. Visit the What's new in the Darwin release page for a summary of changes from the previous Curie release. Documentation for the Curie release can be accessed using the dropdown menu on the top left of the page.

Using Genius

Genius starts with building a probabilistic model. A probabilistic model is a simplified representation of the real world expressed in the language of probability. This model is created from a combination of your own domain knowledge and available data.

The probabilistic model, represented as a factor graph, is then used by a Genius agent to perform various reasoning tasks. These reasoning tasks may optionally include the selection of actions for decision-making purposes with active inference which may be used by the Genius agent to control its environment to achieve some user-specified goal.

There are three primary ways to interact with Genius.

• The Genius model editor allows you to rapidly create and edit probabilistic models in a graphical interface to send to a Genius agent for inference or action selection.

• The API allows you to build apps that interact with Genius through FastAPI calls.

• The Python SDK allows you to easily build models and interact with Genius agents using a Python library.

Features

Genius comes powered with a number of features including:

• Probabilistic inference

• Parameter learning

• Active inference

Where to start

Visit the Is Genius right for me? page to see if your dataset is a good fit for modeling with Genius. If so, proceed to the Genius model editor or Python SDK pages to learn how to use Genius.

If you are new to probabilistic modeling, visit the tutorials listed in the Knowledge Center. These tutorials will provide an overview the fundamental terminology and principles that underlie Genius technology.

Next, learn how to build models from scratch or see Genius in action in the Examples pages which show how to solve various modeling problems with the Genius model editor or Python SDK.

Be sure to visit the references and resource sections for additional information.