Guest editorial of the special issue on intelligent virtual agents

We are pleased to present this special issue on Intelligent Virtual Agents. Intelligent Virtual Agents (IVAs) are autonomous, graphically embodied agents in a dynamic social environment. They are capable of real-time perception, cognition and action, which allows to them interact autonomously and intelligently with the environment, other IVAs, and especially with human users. Research on IVAs now encompasses a wide range of disciplines; artificial intelligence, autonomous agents, robotics, cognitive and social psychology, linguistics and conversation models, affective computing, human-computer interaction, sociology and numerous application domains.This issue contains updated and extended versions of selected papers from last year’s IVA 08 conference, held in Tokyo. They represent the state of the art in the very interdisciplinary research that needs to be carried out to build IVAs. Ito et al. present a decision-theoretic framework for belief maintenance and decisionmaking. It is applied to the human capacity for self-deception, by modeling processes for determining a desired belief state, the biasing of internal beliefs towards this desired belief state, and the decision-making based upon the integrated biases. Si et al. employ a multi-agent decision-theoretic framework to computationally model the cognitive appraisal of events in the environment. Based on the agents’ recursive beliefs about self and others, emotional experience is linked to the interpretations of these events, which allows for simulating “real emotions” in virtual characters and for predicting the human user’s emotional experience. Becker & Wachsmuth introduce an approach to combine primary and secondary emotions to a coherent affective state and emotional expressions of a virtual agent. Primary emotions reflect the continuous progression of bodily feelings, while secondary emotions reflect the cognitive appraisal of events in the light of current experiences and expectations. An evaluation study in a card playing game scenario shows that the inclusion of secondary emotions affects the user’s impression of the agent. Moving into the modeling of interactional behavior, the paper by Lance & Marsella considers expressive gaze shifts as a signal to enhance the believability of a simulated character. Such gaze shifts can be performed in many different ways, each of which can potentially