H. Chad Lane
This chapter is from the forthcoming The Oxford Handbook of Affective Computing edited by Rafael Calvo, Sidney K. D'Mello, Jonathan Gratch, and Arvid Kappas. Institutions of informal learning often seek to influence interest, attitudes, and feelings about the topics they address. The quality of visitor experiences depend on many factors, including the content of the exhibits, the ability of exhibits to promote sustained engagement, and the nature of the conversations visitors have with each other and staff. This chapter discusses the role of emotions during informal learning experiences and how affect-aware technologies could be used to enhance cognitive and affective outcomes. Four potential application areas are presented: (1) automation of evaluation tasks for informal learning, (2) sparking visitors’ interest and magnifying the attracting power of exhibits, (3) deepening engagement during learning activities, and (4) promoting productive conversational behaviors in groups as well as single visitors (with virtual agents). Key challenges ahead include the development of robust detection algorithms, addressing privacy concerns, tracking visitors beyond single exhibits, and integrating heterogeneous sources of information into useful estimates of visitors’ knowledge, emotions, and goals.
Ryan S.J.d. Baker and Jaclyn Ocumpaugh
This chapter is from the forthcoming The Oxford Handbook of Affective Computing edited by Rafael Calvo, Sidney K. D'Mello, Jonathan Gratch, and Arvid Kappas. In recent years, the essential role that affect plays during learning has received greater attention. Accordingly, there has been increasing interest in developing affect-sensitive learning systems that can infer student affect and respond appropriately to it. In many domains, researchers have leveraged physical sensors to make substantial progress in affect detection, but this approach can be challenging in educational settings, where physical sensors can be infeasible. As such, there has been increasing interest in developing detectors of student affect that operate solely on data from the interaction (within the user interface) between the student and the computer. This chapter reviews recent research in this area, discussing approaches used both for collecting training labels of affect and approaches for predicting those affect labels from features of the student interaction. The authors also discuss some of the methodological insights that are emerging from this area of research.