Affectively Framework: Towards Human-like Affect-Based Agents

TL;DR

The Affectively Framework introduces a set of Open-AI Gym environments that incorporate human affect models into reinforcement learning, enabling more human-like virtual agents by integrating affect into observations.

Contribution

It is the first framework to embed affect models into reinforcement learning environments, enhancing the development of human-like affect-based agents.

Findings

Baseline experiments validate the framework's effectiveness.

Affect integration improves agent interaction quality.

The framework supports diverse game environments.

Abstract

Game environments offer a unique opportunity for training virtual agents due to their interactive nature, which provides diverse play traces and affect labels. Despite their potential, no reinforcement learning framework incorporates human affect models as part of their observation space or reward mechanism. To address this, we present the \emph{Affectively Framework}, a set of Open-AI Gym environments that integrate affect as part of the observation space. This paper introduces the framework and its three game environments and provides baseline experiments to validate its effectiveness and potential.