Abstract

Previous active inference accounts of emotion translate fluctuations in freeenergy to a sense of emotion, mainly focusing on valence. However, in affectivescience, emotions are often represented as multi-dimensional. In this paper, wepropose to adopt a Circumplex Model of emotion by mapping emotions into atwo-dimensional spectrum of valence and arousal. We show how one can derive avalence and arousal signal from an agent's expected free energy, relatingarousal to the entropy of posterior beliefs and valence to utility lessexpected utility. Under this formulation, we simulate artificial agents engagedin a search task. We show that the manipulation of priors and object presenceresults in commonsense variability in emotional states.

 

Quick Read (beta)