14:00   Oral Session 6-KzZ – Models of Emotions 2
Chair: Jeff Cohn
14:00
25 mins 		Assessing the validity of appraisal-based models of emotion
Jonathan Gratch, Stacy Marsella, Ning Wang, Brooke Stankovic
Abstract: We describe the results of an empirical study comparing the performance of competing computational models of emotion. We review criteria under which such models could be evaluated, here focusing on models’ ability to accurately predict human emotional responses in naturalistic emotion-eliciting situations. We survey proposed models of emotion and highlight their contradictory design assumptions. We then contrast these assumptions with behavior of human subjects. The results find clear differences in these models ability to accurately forecast human emotional responses, and provide guidance on how to develop more accurate models of human emotion.
14:25
25 mins 		Same or Different? Recollection Of or Empathizing With an Emotional Event From The Perspective of Appraisal Models
Gert-Jan de Vries, Paul M.C. Lemmens, Dirk Brokken
Abstract: The same stimulus can evoke different emotions for different individuals. Incorporating personalized construal of stimuli is how appraisal models differ from dimensional models of emotion. Scherer formulated a model of the cognitive antecedents of emotions and analyzed recollections of events and emotions that his participants provided. In the present study, we were interested whether Scherer’s appraisal model also applies to a situation in which participants have to empathize with a photo by building a story around the event depicted. Our results show that recollecting a personal emotional event and empathizing with a photo seem to be similar processes. Results of machine-learning techniques, however, show lacking performance of classifications to express discrete emotion labels in terms of appraisal data.
14:50
25 mins 		A socio-emotional model of impoliteness for Non-Player Characters
Sabrina Campano, Nicolas Sabouret
Abstract: Due to its important role in dialogue, politeness has been widely studied both from a theoretical point of view and at a practical level in virtual agents. However, few attention has been given to impoliteness, even though virtual characters, such as Non Player Characters in video games, should be able to use it to improve their believability. In this paper, we describe a model for utterance selection based on impoliteness that takes into account emotions, personality, and social relations.
15:15
25 mins 		Assessing the validity of a computational model of emotional coping
Stacy Marsella, Jonathan Gratch, Ning Wang, Brooke Stankovic
Abstract: In this paper we describe the results of a rigorous empirical study evaluating the coping responses of a computational model of emotion. We discuss three key kinds of coping, Wishful Thinking, Resignation and Distancing, and relate these coping responses to related work in the attitude change literature. We discuss the EMA computational model of emotion and identify several hypotheses it makes of these three coping processes. We assess these hypotheses against the behavior of human subjects playing a competitive board game, using monetary gains and losses to induce emotion and coping. We indexed subject’s appraisals, emotional state and coping responses at key points throughout a game, revealing a coherent pattern in the dynamic relationship between these factors. The results clearly support several of the hypotheses but also identify (a) extensions to how EMA models Wishful Thinking as well as (b) individual differences in subject’s coping responses.
15:40
25 mins 		Evaluation of a Computational Model of Surprise Arousal in Narratives
Byung-Chull Bae, R. Michael Young
Abstract: In this paper we analyze the effectiveness of a planning-based computational model of surprise arousal previously developed by Bae and Young [1]. This evaluation was performed by measuring the effect of stories produced by this method on several factors contributing to surprise in narratives (e.g., expectation failures, the importance of story events, the emotional valence of a reader, and resolution) and their relation to story interest. To measure the ratings of a reader’s surprise and story interest, we conducted a pilot study, where three different discourse types – Chronological, Flashback, and Flashback with Foreshadowing – were compared. A one-way ANOVA and a priori pair-wise comparisons confirmed that the mean ratings of surprise and interestingness increased at the .01 and .05 level, respectively. While the number of subjects in this study was relatively small, the significance of our results argues for support to the claim that the computational model we evaluated can effectively manipulate readers’ sense of surprise in their experience of a narrative.