Behaviors that emerge from emotion and cognition

Abstract

This paper describes the implementation and evaluation of a framework for modeling emotions in complex, decision-making agents. Sponsored by U.S. Army Research Institute (ARI), the objective of this research is to make the decision-making process of complex agents less predictable and more realistic, by incorporating emotional factors that affect humans. In tune with modern theories of emotions, we regard emotions essentially as subconscious signals and evaluations that inform, modify, and receive feedback from a variety of sources including higher cognitive processes and the sensorimotor system. Thus, our work explicitly distinguishes the subconscious processes (in a connectionist implementation) and the decision making that is subject to emotional influences (in a symbolic cognitive architecture).It is our position that "emotional states" are emergent patterns of interaction between decision-making knowledge and these emotional signal systems. To this end, we have adopted an approach that promotes the emergence of behavior as a result of complex interactions between factors affecting emotions, integrated in the connectionist-style model, and factors affecting decision making, represented in the symbolic model. This paper presents the implementation of emotions architecture and explains how we evaluated the system. This includes a description of the behaviors we used in our prototype, the design of our experiments, a representative set of behavior patterns that emerged as a result of exercising our model over the design space, and our project's lessons learned

Invited Commentaries

“…As each choice in a model that is not related to the main hypothesis needs to be argued for, the best choice is to not have to make one. For example, if you are interested in the emergent nature of behavior, emotion and cognition or the specific relations that might exist between decision making and affect, then build upon an existing cognitive architecture to avoid unnecessary assumptions (Hogewoning, Broekens, Eggermont, & Bovenkamp, 2007;Henninger, Jones, & Chown, 2003;Belavkin, 2001). Each assumption introduces the need to model that assumption and as mentioned earlier, the devil is in the computational details: what is a belief in my framework, what is a desire, what is time, how do I traverse an appraisal tree, how heavily should different appraisal components be activated?…”

Modeling the Experience of Emotion

“…Useful means that the model serves some goal. In virtual training systems (Gratch & Marsella, 2004;Henninger, Jones, & Chown, 2003), the goal is to enhance believability of the agent in order to increase the effect of the training on the trainees. In tutor and support systems (Bickmore & Picard, 2005;Graesser, Chipman, Haynes, & Olney, 2005;Heylen, Nijholt, Akker, & Vissers, 2003), the goal is to enhance the relation between the user and the system such that the system is used more often, longer and/or more effectively.…”

A Software Tool for Estimating Driver Effectiveness during Evacuations

“…Although there have been many advances in human performance modeling over the last decade, the focus has been on the prediction of human-system dynamics in typical system operation rather than focusing on the underlying dynamics behind and the consequences of disrupted systems. Of course, there is good reason for this -if models cannot predict typical performance well, they will not be able to support the more challenging demands of predicting and dealing with atypical performance (Henninger, et al, 2002). In short, we must "walk" with human performance models before we can expect to "run.…”