Impact of Prosody and Facial Expressions on Interactive Narration with a Child Ovidiu S , erban Normandie Univ, INSA Rouen Normandie, LITIS 76800 Saint-Étienne-du-Rouvray France Mukesh Barange Normandie Univ, INSA Rouen Normandie, LITIS 76800 Saint-Étienne-du-Rouvray France Sahba Zojaji Normandie Univ, INSA Rouen Normandie, LITIS 76800 Saint-Étienne-du-Rouvray France Alexandre Pauchet Normandie Univ, INSA Rouen Normandie, LITIS 76800 Saint-Étienne-du-Rouvray France Adeline Richard Normandie Univ, UNIROUEN, PSY-NCA 76000 Rouen, France Emilie Chanoni Normandie Univ, UNIROUEN, PSY-NCA 76000 Rouen, France ABSTRACT Intelligent Virtual Agents are suitable means for interactive sto- rytelling for children. The engagement level of child interaction with virtual agents is a challenging issue in this area. However, the characteristics of child-agent interaction received moderate to little attention in scientifc studies whereas such knowledge may be crucial to design specifc applications. This article proposes a Wizard of Oz platform for interactive narration. An experimental study in the context of interactive story- telling exploiting this platform is presented to evaluate the impact of agent prosody and facial expressions on child participation during storytelling. The results show that the use of the virtual agent with prosody and facial expression modalities improves the engagement of children in interaction during the narrative sessions. CCS CONCEPTS · Human-centered computing → User studies; Human com- puter interaction (HCI); KEYWORDS Child-Agent Interaction, Interactive Narration, Prosody, Facial Ex- pression, WoZ ACM Reference Format: Ovidiu S , erban, Mukesh Barange, Sahba Zojaji, Alexandre Pauchet, Adeline Richard, and Emilie Chanoni. 2018. Impact of Prosody and Facial Expressions on Interactive Narration with a Child. In Proceedings of Workshop Afect, Compagnon Artifciel, Interaction (WACAI’18). Ile de Porquerolles, France , 8 pages. Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for proft or commercial advantage and that copies bear this notice and the full citation on the frst page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specifc permission and/or a fee. Request permissions from permissions@acm.org. WACAI’18, June 13ś15, 2018, Ile de Porquerolles, France © 2018 Association for Computing Machinery. 1 INTRODUCTION Designing a virtual environment, where the participants can inter- act without any difculty, is very challenging. Particularly, introduc- ing an autonomous dialogue-based virtual character (or Embodied Conversational Agent (ECA) [8]) increases the expectations of the human participants, up to the point where they can be disappointed by the agent’s capabilities [24]. Building such an environment is even more difcult when young children are involved, since they are still developing their linguistic and interaction competencies. Providing children with a non-disturbing environment which ofers natural reactions from the included ECAs, becomes critical. Among the various applications of ECAs, interactive storytelling is a growing scientifc feld [12]. It includes situations that corre- sponds from a reproduction of the familiar parent-child narration up to a new user experience with story generation according to the interaction. Nevertheless, interactive narration aims at improving the user’s immersion, pleasure, feeling of control, believability of the virtual characters and interaction engagement [46]. A few experiments exist regarding interactive storytelling with children (e.g. [31, 39, 46]), but unfortunately they do not enable to determine and characterize standard data in child-agent interaction, such as the average response time (latency) of a child interacting with an ECA. In this article, we propose an interactive environment with a virtual character, centered around a familiar story telling activity so that the children feel comfortable. As the dialogue component of ECAs remain a technical difculty [18, 44], we propose a Wizard of Oz (WoZ) platform in order to constitute desired virtual environment and interaction scenario to evaluate preliminary components. Thus, our environment is based on the WoZ paradigm, so that the collected data expresses what can be expected of ‘natural’ interaction. We aim at answering the following research question: Do diferent modalities of virtual agent have any impact on child-agent interaction? This question is stud- ied regarding interaction engagement. We therefore propose an experimental study to compare the efects of diferent agent modal- ities (e.g. prosody and facial expression) on the child ↔ avatar 1 interaction during interactive storytelling sessions. 1 In the following, ‘avatar ’ refers to a virtual character driven during a WoZ experi- ment. A ‘virtual character ’ can be either an avatar or an ECA.