Sorrel: A simple and flexible framework for multi-agent reinforcement learning

TL;DR

Sorrel is a user-friendly Python framework designed to facilitate the creation and testing of multi-agent reinforcement learning environments, emphasizing simplicity and psychological intuitiveness for social science research.

Contribution

It introduces a flexible, accessible interface for multi-agent RL environments, bridging reinforcement learning and social science research.

Findings

Provides a new tool for social scientists to explore group dynamics.

Simplifies the development of multi-agent RL environments.

Enhances understanding of social interactions in learning processes.

Abstract

We introduce Sorrel (https://github.com/social-ai-uoft/sorrel), a simple Python interface for generating and testing new multi-agent reinforcement learning environments. This interface places a high degree of emphasis on simplicity and accessibility, and uses a more psychologically intuitive structure for the basic agent-environment loop, making it a useful tool for social scientists to investigate how learning and social interaction leads to the development and change of group dynamics. In this short paper, we outline the basic design philosophy and features of Sorrel.