Gaze, conversational agents and face-to-face communication

Abstract

HAL is a multidisciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Social Behavior Modeling Based on Incremental Discrete Hidden Markov Models

“…We used the dataset of Bailly et al [6], who collected speech and gaze data from dyads playing a speech game via a computer-mediated communication system that enabled eye contact and dual eye tracking. The experimental setting is shown in Fig.…”

“…2. Mediated face-to-face conversation [6]. Top: People sit in two different rooms and dialog through couples of cameras, screens, microphones and loudspeakers.…”

“…Our modeling framework and its current implementation are introduced in section 3. Section 4 illustrates its modeling performance using speech and gaze data collected in a previous experiment [6] and shows the results. Finally, discussions and our conclusion are presented in section 5.…”

Measuring the differences between human-human and human-machine dialogs

“…They have also been used for education and training, particularly in improving phonetic and linguistic skills: assistance and guidance to F18 aircraft personnel during maintenance tasks (Bohus and Rudnicky, 2002), dialog applications for computer-aided speech therapy with different language pathologies (Vaquero et al, 2006). Finally, one of the most demanding applications for fully natural and understandable dialogs, are embodied conversational agents and companions (Brahnam, 2009;Bailly et al, 2010).…”

Towards the Use of Dialog Systems to Facilitate Inclusive Education

“…Researchers have also investigated user preferences for the agent expertise. Findings suggest that in general, similarities in competency between an agent and learner have positive impacts on the learners' affective attainments, for example, academically strong students showed higher self-efficacy beliefs in a task after working with a high-competency agent, while academically weak students showed higher self-efficacy after working with a low-competency agent (Kim, 2007 (Bailly et al, 2010) (Edlund et al, 2008), involving them in a deeper and intimate interaction, it is difficult to communicate with these agents whenever needed (i.e. when the user is not in front of a computer but he/she has the need to get suggestions and advices).…”