Affect plays a key role in learning and to achieve enhanced learning outcome, it is crucial that the active engagement and motivation of the learner be sustained through this continuous effective feedback loop between the tutor and the learner.
It is thus essential for tutoring systems to sense the affect of learners, analogous to a personal tutor that continually sense the exhibited affect and adapt the tutoring interactions to sustain the learning motivation.
This is a new method or system for sensing of learner or novice programmer’s frustration and dis-engagement through keystrokes, mouse clicks, head postures, and facial dynamics captured using an array of unobtrusive and ubiquitous sensors and scaffolding of their learning through the generation of sensitive contextual hints and affective prompts.
The obtrusiveness and cost of the sensors are key considerations which will influence users’ acceptance, accuracy, scalability, deployment, and use of the tutoring system. Thus, an array of unobtrusive and ubiquitous sensors are proposed here to capture keystrokes, mouse clicks, head postures and facial dynamics for the inferring and sensing of learner or novice programmer’s frustration and engagement.
Advanced machine learning techniques are used here to infer the affective states of learners from the captured sensor readings on a 30 seconds detection granularity.
In the envisaged online tutoring environment, the learners’ frustration and engagement are monitored on a granular basis for adaptation of tutoring strategies to optimize their learning. Attaining prompt tutoring responses to the fast-changing nature of learners’ affective states is instrumental to optimizing the learning of students in an online tutoring context.
This solution can potentially be deployed in educational, training institutions and Massive Open Online Course (MOOC) providers as well as in special needs education.
With the increasing importance of life-long learning, having a self-directed learning tool that can make learning as effective as face-to-face learning would be helpful to learners. Moreover, with the use of unobtrusive sensors, the tutoring system is easily accessible and learning can take place anywhere, anytime.
Potential customers will benefit from a readily deployable module which augments their tutoring system capability by both capturing and inferring the affective states of their learners. The affective module would be able to infer valuable affective insights and this will help customers to understand the learning patterns of their learners. The module can potentially be extended to recommend adequate tutoring strategies to optimize learners’ learning