Lifelike interface agents (e.g. talking avatars) have been increasingly used in human-computer interaction applications. In this work, we quantitatively analyze how human perception is affected by audio-head motion characteristics of talking avatars. Specifically, we quantify the correlation between perceptual user ratings (obtained via user study) and joint audio-head motion features as well as head motion patterns in the frequency-domain. Our quantitative analysis results clearly show that the correlation coefficient between the pitch of speech signals (but not the RMS energy of speech signals) and head motions is approximately linearly proportional to the perceptual user rating, and a larger proportion of high frequency signals in talking avatar head movements tends to degrade the user perception in terms of naturalness. Author Keywords Perceptual modeling, quantitative analysis, head motion, and

This study investigates effectiveness of a local high-fidelity 3D facial avatar for a global audience by observing how US and International student groups differed in identifying subjects and perceiving emotions while viewing nonverbal high-fidelity 3D facial avatar animations embedded with the motion data of three US individuals. To synthesize the animated 3D avatars to convey highly believable facial expressions, a 3D scanned facial model was mapped with high-fidelity motioncapture data of three native US subjects as they spoke designated English sentences with specified emotions. Simple animations in conjunction with actual footage of the subjects speaking during the facial motion-capture sessions were shown several times to both native US and international students in similar settings. After a familiarization process, we showed the students randomly arranged talking avatars without voices and asked them to identify the corresponding identities and emotional types of the subjects whose facial expressions were utilized in the creation of the avatars, and to rate their confidence in their selections. We found that the US group had higher success rates in subject identification, although the related difference in confidence ratings between two groups was not significant. The differences in the success rates and confidence ratings on the perception of emotion between the two groups were not significant either. The results of our

Because ethos is an unavoidable component of dialogue and forms the basis for believing and being persuaded by another's speech, it is an important topic for AI researchers. This paper examines the concept of ethos, especially Aristotle's notions of situated and invented ethos, as it functions in oral and written discourse and then explores what happens to ethos in computer-mediated human-to-human and human-to-machine discourse. The paper draws a number of conclusions that may be of value to researchers in these fields. In particular, it argues that the rhetorical concept of ethos furnishes a broader theoretical framework for understanding design and ethical issues involved in agent credibility than does the artistic notion of believability. The paper concludes by suggesting some nonartistic methods for making agents more credible within the framework of situated ethos.

Community designers can draw from theories of commitment to make design decisions that influence whether and how people will become committed to a community. Commitment is harder to achieve than a flow (or trickle) of visitors, but for most online communities, commitment is crucial. Committed members work harder, say more, do more, and stick with a community after it becomes established. They care enough to help with community activities and to sustain the group through problems. Committed members are those most likely to provide the content that others value, such as answers to people’s questions in technical and health support

Simulated humans in computer interfaces are increasingly taking on roles that were once reserved for real humans. The presentation of simulated humans is affected by their appearance, motion quality, and interactivity. These presentational factors can influence the decisions of those who interact with them. This is of concern to interface designers and users alike, because these decisions often have moral and ethical consequences. However, the impact of presentational factors on decisions in ethical dilemmas has not been explored. This study is intended as a first effort toward filling this gap. In a between-groups experiment, a female character presented participants with an ethical dilemma. The character’s human photorealism and motion quality were varied to generate four stimulus conditions: real human versus computer-generated character � fluid versus jerky movement. The results indicate that the stimulus condition had no significant effect on female participants, while male participants were significantly more likely to rule against the character when her visual appearance was computer generated and her movements were jerky. 1

In a project designed to test how viewers evaluate and choose avatars to represent them online, viewers (N = 261) evaluated 92 potential avatars that varied along three dimensions – gender of the image, anthropomorphism (humanness), and realism (veridicality). Consistent with the predictions of information processing theory, results indicated that anthropomorphic qualities increased perceived realism and subsequent homophily judgments. As predicted by the uncanny valley hypothesis, viewers preferred avatars with levels of realism and anthropomorphism matched and were more critical of imperfections of avatars with higher levels of anthropomorphism. Image gender and anthropomorphism were found to produce a combination of positive and negative effects on competence ratings and selection of avatars. The theoretical and practical implications of these visible feature effects of avatars are discussed. 1.

Abstract—Mediation is an important paradigm for dispute resolution. It can lead to “win-win ” situations and benefit all parties. Yet, when people negotiate they demonstrate bounded rationality in their actions and diversity in their behaviors. This increases the difficulty to design automated mediators. To be successful, the mediator must take this into account and propose solutions deemed relevant, otherwise it will lose the focus and trust of the negotiators. To tackle these difficulties we present AniMed * – an automated animated mediator, incorporated with a novel proposal generation strategy, aimed to increase the social benefit of the negotiating parties. To validate the benefits of using AniMed * in negotiations, experiments were conducted with 130 people negotiating with each other. The results demonstrate the significant increase both in the social welfare and the individual utilities of both parties, compared to negotiations in which another version of reduced functionalities of AniMed * was used, as well as negotiations involving another stateof-the-art automated mediator or no mediator. Index Terms—bilateral negotiations, automated mediation, incomplete information.

Prior research has investigated the effect of interactive social agents presented on computer screens or embodied in robots. Much of this research has been pursued in labs and brief field studies. Comparatively little is known about social agents embedded in the workplace, where employees have repeated interactions with the agent, alone and with others. We designed a social robot snack delivery service for a workplace, and evaluated the service over four months allowing each employee to use it for two months. We report on how employees responded to the robot and the service over repeated encounters. Employees attached different social roles to the robot beyond a delivery person as they incorporated the robot’s visit into their workplace routines. Beyond one-on-one interaction, the robot created a ripple effect in the workplace, triggering new behaviors among employees, including politeness, protection of the robot, mimicry, social comparison, and even jealousy. We discuss the implications of these ripple effects for designing services incorporating social agents. Author Keywords Social agent, human-robot interaction, service design,