Organization Important Dates Keynote Speakers Program Registration Location Contact Us


Keynote Speakers



Helen Pain

University of Edinburgh, UK


Dr Helen Pain is a Senior Lecturer in ICCS, University of Edinburgh, with degrees in Psychology and Artificial Intelligence (AI) and Education. Her current research interests include modelling social intelligence in pedagogical agents; affective and cognitive modelling of learners; managing educational dialogue; user-centred design and evaluation of interfaces, and empirical methodology. She researches and teaches in the areas of AI in Education (AI-Ed), and User Modelling, and has contributed to numerous conferences, reviewed papers, held grants and published in these areas. She is currently involved in the EU project LeActiveMath and the EPSRC project STANDUP. Dr Pain is past-president of the International AI-Ed Society.


Affect in One-to-One Tutoring

Helen Pain and Kaska Porayska-Pomsta
ICCS/HRCR, University of Edinburgh, 2 Buccleuch Place, Edinburgh, Scotland, EH8 9LW, UK
[email protected]
[email protected]

It is well known that human tutors take into account both the student’s knowledge and understanding of what is being taught, in addition to considering the emotional and motivational state of the student. However, there are many gaps in our understanding of the relationship between cognition and affect in tutoring. We have some insight into how human tutors infer student’s cognitive and affective states, and current research has attempted to apply this knowledge to the inference of such states by computer tutors. There is ongoing research on how human tutors use their knowledge of student’s states in their decisions and actions, and how we might use such research to inform the design of computer tutors.

In this talk we will consider what is currently known about how human tutors infer emotion and motivation in students, whether and how they act of these inferences, and how this relates to the student’s cognitive state. We will describe methods that may be used to infer affect, and how these might be adapted for use in Human-Computer educational interactions, providing illustrations from our current work in which we explore how human tutors diagnose and manipulate affect in situational contexts, what affective states may be relevant to tutoring, and how we can model them formally. In particular we will raise questions about how fine grained the diagnosis of affective states needs to be, and whether in fact we need to diagnose (or know how to act on) mid-range affective states -or whether we might consider acting only in response to extreme emotional and motivational state values.


Research Center for Science and Technology for Learning
National Central University