​Tutorial Dialogues and Spoken Dialogue Systems.

The special session focuses on methods, problems and challenges concerning issues emerging from the application of spoken and multimodal interaction technology into various educational settings. It will survey issues related to processing of spoken language in face-to-face interactions, modelling of the teacher-student interaction in MOOC-like environments, and evaluating tutorial dialogue systems from the point of view of natural interaction, technological robustness, and learning outcome. 

 

An important aspect of effective communication is the use of multimodal signals in giving and eliciting feedback among the interaction partners, concerning their emotional state as well as their understanding of the current state of the interaction and the learning content. For educational applications, there is a particular need to identify and model hand gestures, body postures, and facial expressions, that are used in the estimation of various speaker states related to understanding and emotional states. Besides verbal communication, the existing speech and multimodal processing technologies must be adapted and new modules added to interactive systems, so as to improve their often rigid and simple communication strategies, and to cater for the detection and generation of social

signals in educational interactions. These experimental models can then be used for designing and developing tutorial dialogue systems that interact with learners using natural language in an intuitive manner. The special session will provide an overview of these issues and deepen our understanding of how different dialogue management techniques compare against each other in modelling feedback and turn-taking strategies to ground information and to create shared understanding.

 

The special session also encourages interdisciplinary research where education experts, social scientists, dialogue researchers, engineers, and communication technologists can work together in the design and construction of interactive educational applications, and experimenting with novel technological possibilities such as human-robot interactions for the fun and engaging interactions. Moreover, education and interaction are two fields where much intercultural communication takes place, as both the teaching and learning processes will take various preferred forms depending on the cultural settings and practises. Models for these kind of aspects across languages and cultures are pertinent to take into account in the design of educational systems for a wide range of students and educational situations.

 

The relation between the teacher and the learner is also important, and it becomes a relevant design concept when considering humanoid robots as teachers or peer-like learning companions. Such agents can show high-level spoken language technology capacities and be able to recognise emotion, display aspects of a distinct personality (or personalities), and to interact with a student over a long period, thus learning about their needs and interests, providing entertaining, consoling, and so on. The importance 

of good cognitive underpinnings is also eessential. Abilities to determine what a student has understood, when to change topics, and how to assess understanding in a robust manner are a part of these concerns. Detection and tracking of motivation and engagement reveal much about the quality of the learning situation.

 

The special session is thus especially related to the special focus of the conference, "Celebrating the Diversity of Spoken Languages", as it directly addresses the issues related to the learning and use of spoken languages for communication and learning, as well as explores the use of spoken language technology in the development of various tutorial applications.