Learning and emotions
Automatic inference of affect
This work explores ways of augmenting learning environments by incorporating emotional intelligence in them. It is motivated by the importance of emotions in the learning process and the fact that as yet machine-learner interactions do not address this crucial link. In naturalistic settings, the availability of several channels of communication facilities the constant monitoring necessary for an interactive, effective and flexible learning experience. As learning with computers is essentially self-paced, assessing the learner's experience can give information which can be helpful in adapting the tutorial interaction and strategy. Such a responsive interface can also alleviate fears of isolation in learners and facilitate learning at an optimal level. The idea is to emulate the social dynamics of expert human mentoring using appropriate affective diagnoses.
This research is specifically focused on examining the utility of facial affect analysis to model the affective state of a learner in a one-on-one learning setting. Though facial affect analysis using posed or acted data has been studied in great detail, research using naturalistic data still remains a challenging problem. Since the context of this research is a real application environment, it is based on naturalistic data. Of interest to this study are the social responses to such affect-sensitive socially intelligent machines, and how these might influence subsequent interaction and behaviour.
- Natural Affect Data - Collection & Annotation in a Learning Context
Description of naturalistic data collection and annotation process
- Perception of Emotional Expressions in Different Representations Using Facial Feature Points
Experiment analysing the emotional information encoded by automatically tracked facial feature points
- Intentional affect: An alternative notion of affective interaction with a machine
An alternative design based approach to measure affective experience
- Measuring Affect in Learning – Motivation and Methods
Affect inference in learning environments