Mind-, Body-and Emotion-Reading

A core requirement of human-like intelligent agents is the ability to create and dynamically maintain mental models of other agents. We suggest that such modeling capabilities must extend beyond traditional notions of “mindreading” to include inferring unobservable features of other agents’ bodies and emotions as well. We describe the modeling of agents endowed with a mind, body and emotions within the OntoAgent knowledge environment and discuss how the same ontologically-grounded, mentalist models that support the simulation of agents also permit those agents to carry out mind-, bodyand emotion-reading of their human and artificial collaborators.