APReL: A Library for Active Preference-based Reward Learning Algorithms

TL;DR

APReL is an open-source library that facilitates experimentation and development of active preference-based reward learning algorithms for human-robot interaction.

Contribution

It introduces a comprehensive library that consolidates existing algorithms and tools for reward learning, simplifying research and development in this area.

Findings

Provides a modular framework for preference-based reward learning

Enables easy comparison of different algorithms

Supports development of new reward learning techniques

Abstract

Reward learning is a fundamental problem in human-robot interaction to have robots that operate in alignment with what their human user wants. Many preference-based learning algorithms and active querying techniques have been proposed as a solution to this problem. In this paper, we present APReL, a library for active preference-based reward learning algorithms, which enable researchers and practitioners to experiment with the existing techniques and easily develop their own algorithms for various modules of the problem. APReL is available at https://github.com/Stanford-ILIAD/APReL.