Recent psychological theories of emotion have explicitly focused on modeling the multiplicity of levels of the human emotion system. This increased interest in creating multilevel theories of emotion matches similar efforts in the area of cognitive architectures. In this paper we report our initial work toward integrating a psychological multilevel model of emotions with Brahms, a multi-agent system that is used to model and simulate work practice.
Appraisal theory, a functional approach to understanding emotion elicitation is described. Three distinct classes of appraisal models are reviewed: structural - which describe the cognitive contents of appraisal and how those contents map onto the elicitation of various distinct emotions; procedural - which describe the cognitive processes underlying appraisal; and relational - which describe how both person and situation information is combined in producing specific appraisal outcomes. A theoretical example of each class of model is described, and the state of the empirical literature addressing such models is reviewed. The relevance of the general theoretical approach, and of the three types of appraisal models, to developing architectures for modeling emotion are discussed.
Ferreira, Nuno (INESC-ID and Instituto Superior Tecnico) | Mascarenhas, Samuel (INESC-ID and Instituto Superior Tecnico) | Paiva, Ana (INESC-ID and Instituto Superior Tecnico) | Tosto, Gennaro Di ( Utrecht University ) | Dignum, Frank (Utrecht University) | Breen, John Mc ( Wageningen University ) | Degens, Nick ( Wageningen University ) | Hofstede, Gert Jan ( Wageningen University ) | Andrighetto, Giulia ( ISTC-CNR ) | Conte, Rosaria (ISTC-CNR)
Emotional synthetic characters are able to evaluate (appraise) events as positive or negative with their emotional states being triggered by several factors. Currently, the vast majority of models for appraisal in synthetic characters consider factors related to the goals and preferences of the characters. We argue that appraisals that only take into consideration these "personal" factors are incomplete as other more social factors, such as the normative and the social context, including in-group and out-group relations, should be considered as well. Without them, moral emotions such as shame cannot be appraised, limiting the believability of the characters in certain situations. We present a model for the appraisal of characters' actions that evaluates whether actions by in-group and out-group members which conform, or not, to social norms generate different emotions depending on the social relations between the characters. The model was then implemented in an architecture for virtual agents and evaluated with humans. Results suggest that the emotions generated by our model are perceived by the participants, taking into account the social context and that participants experienced very similar emotions, both in type and intensity, to the emotions appraised and generated by the characters.
In this paper we present a novel approach to a grounded synthesis of emotional appraisal, based on a multicausal model of the appraisal process. We investigate the functional nature of emotion by implementing a robotic model in a predator/prey scenario which is able to discriminate and anticipate outcomes through emotional appraisal. The robots evolve to react in apparently emotional ways, showing how the functionality of emotion can emerge naturally. We demonstrate through this implementation the value of emotion appraisal as a form of anticipation. This supports the view that emotional behavior can often be seen as an effective alternative to rational cognition. Our effort here is to build a model that can be simultaneously seen as belonging to both NCS and more classical theorizing based on cognitions and representations, understandable both mechanically and subjectively from a human standpoint.
Appraisal processes provide an affective assessment of an agent's current situation, in light of its needs and goals. This paper describes a computational model of the appraisal process, implemented within the broader context of a cognitive agent architecture. A particular focus here is on modeling the interacting influences of states and traits on perception and cognition, including their effects on the appraisal process itself. These effects are modeled by manipulating a series of architecture parameters, such as the speed and processing capacity of the individual modules. The paper presents results of an evaluation experiment modeling the behavior of three types of agents: 'normal', 'anxious', and'aggressive'. The appraisal model generated different affective appraisals of the same set of external circumstances for the different agent types, resulting in distinct emotions, and eventually leading to observable differences in behavior. The paper concludes with a brief discussion of some of the issues encountered during the appraisal model development.