A Review of Emotion Regulation in Intelligent Tutoring Systems.
Mehdi Malekzadeh, Mumtaz Begum Mustafa, Adel Lahsasna · Educational Technology & Society · 2015
Introduction A common view of emotions is that they are generated as a results of human's judgment about the world and initiated by individual's appraisal in response to and interaction with stimulus, such as material that the individual is learning (Desmet, 2002; Lazarus, 1991). Recent findings in neuroscience and psychology found that emotions are widely related to cognition, influencing various behavioural and cognitive processes, such as attention, long-term memorizing, decision-making, and so on (Ahn & Picard, 2005). Researches on emotion and learning suggest that positive emotions (affects) have a vital influence on various cognitive processes relevant for learning, such as information processing, communication processing, decision-making processing, negotiation processing, category sorting tasks, and creative problem-solving processes (Erez & Isen, 2002). Positive emotions promote higher cognitive flexibility and allow the learner to discover new ideas and possibilities. In addition, as a function of positive emotion, cognitive processes may be more flexible that result in greater creativity and improved problemsolving (Isen et al., 1987). Emotion also influences memory, where positive emotional state improved recall and it served as effective retrieval cues for long-term memory in many experiments (Isen et al., 1978). Reciprocally, negative emotional states like boredom and frustration have been linked with less use of self-regulation and cognitive strategies for learning as well as increases in disengaged and disturbing behavior during learning in the class (Isen, 2001). Thus, emotions, governed by proper attention, self-regulation and motivational strategies result in positive effects on learning, and lead to better achievement among the learners (Pekrun, Goetz, Titz, & Perry, 2002). In traditional learning environment, a teacher maintains a sympathetic relationship with learners to facilitate the development of positive emotions. For instance, students who feel happy generally perform better than students who feel sad, angry, or scared (Connor & Davidson, 2003). This relationship also exists in a computerized learning environments and researchers of computer science in education field had studied techniques of artificial intelligence to make the educational systems more customized to the emotional state (affective states) of students (Jaques & Vicari, 2007). Intelligent tutoring system (ITS) is a computer-based educational system that provides individualised instructions similar to like a human tutor. Typical ITSs determine how and what to teach a student based on the learner's pedagogical state to enhance learning. As experienced human tutor manages the emotional states of a learner to motivate him or her and to improve the learning process, researchers also have augment the learner model structure in ITSs to determine the emotional state of learners (Neji, Ben Ammar, Alimi, & Gouarderes, 2008). Researchers endow ITSs with the ability to detect learners' unpleasant emotional states (e.g., confusion, frustration, etc.), respond to these states, and generate appropriate tutoring strategies as well as emotional expressions by embodied pedagogical agents. These emotion-sensitive ITSs aspire to narrow the interaction bandwidth between computer tutors and human tutors with the hope that this will lead to an improved user experience and enhanced learning gains (Aghaei Pour, Hussain, AlZoubi, D'Mello, & Calvo, 2010; Klein, Moon, & Picard, 2002). In embedding emotional state reasoning into ITSs and intelligent learning environments, there are two main issues that are faced by the developers. First is determining the emotional states of the target learners, and second is determining factors that causes those states as well as how to respond and regulate negative emotional state (Avramides & Du Boulay, 2009; Du Boulay, Rebolledo Mendez, Luckin, & Martinez-Miron, 2007). …