Welcome to APReL’s documentation!

APReL is a unified Python3 library for active preference-based reward learning methods. It offers a modular framework for experimenting with and implementing preference-based reward learning techniques; which include active querying, multimodal learning, and batch generation methods.

Not sure what preference-based learning means? Read on or check our talk at AI-HRI 2021. You can also take a look at our video which contains a simple example.

Indices and tables