Game Development Reference
In-Depth Information
The student model enables the ITS to under-
stand the information and requirements of each
student (Sucar & Noguez, 2008). In addition to
allowing knowledge of the cognitive state of the
learner, the student model can also identify the
learner's emotional (Conati & Maclaren, 2009) or
motivational (Rebolledo-Mendez, Du Boulay, &
Luckin, 2006) states. Emotion influences learning
and performance and is deeply intertwined with
cognitive and motivational constructs (Pekrun et
al., 2007). Affective gaming arises from the union
of the Affective Computing field (Picard et al.,
2004) with the Digital Games field and focuses
on influencing and identifying the player's emo-
tional state (Sykes, 2006). Emotions can be used
by computers for evaluating the user's experience,
understanding when the emotion is relevant to
the learning experience and when it can be safely
ignored, predicting the learner's behavior and
controlling the game. Making machines capable
of recognizing and expressing emotion promises
to enhance the user's experience, e.g. improved
learning, engagement, motivation and training.
is to acquire large quantities of data that will be
classified through AI mechanisms, e.g. Artificial
Neural Networks (ANNs). In addition, the effects
are modeled through the observations of peers, i.e.
lecturers and students, and trained judges. Hard-
ware may be perceived as intrusive and may be
prone to failure. Also this approach is not suitable
for online learning, since students usually have
to travel to special laboratories for interacting
with these systems (Burlenson & Picard, 2007).
Ekman and Friesen (1978) signaled that there is
universality in facial expressions to convey an
emotional message, and as a result, research has
been undertaken that maps facial expressions to
an affective state.
The Ortony, Clore and Collins (OCC) model
is a cognitive theory of emotion that has often
been applied to inferring emotion from its ori-
gin (Chalfoun, Chaffar, & Frasson, 2006). This
theory explains how emotion arises according to
the interplay of a user's goals with events and a
user's standards and attitudes (Ortony, Clore, &
Collins, 1990). Therefore, this approach reasons
about the potential cognitive antecedents of emo-
tions and infers their existence over time. The
challenges of this approach are, being aware of
the learners' standards and attitudes and assessing
the learner's goals. A hybrid approach predicts
whether the emotion will happen according to
the antecedents' existence. To know if the emo-
tion actually happened, the effects have to be
identified and compared with the prediction. This
hybrid approach inherits the challenges of both
approaches (Conati & Maclaren, 2009). Jaques &
Vicari (2007) attempted to infer emotion from its
origin using the learner's observable behavior. One
challenge of applying the OCC model to inferring
the learner's emotional state is to adapt the theory
to the learning context. Jaques & Vicari (2007)
built two student models describing two kinds
of students: one seeking to master the topics and
another focused purely on performance. Effort
was measured qualitatively and quantitatively
during the learner's interaction. The work of Del
Identifying the Learner's Emotional
and Motivational States
The approaches applied by ITSs to infer or rec-
ognize the learner's emotional state are:
1. Recognizing the physical effects of emotion
2. Predicting emotions from their origin
3. A hybrid approach using the previous two
4. Reasoning about the learner's observable
For recognizing the physical effects of emotion,
cameras, microphones and sensors are employed
to collect information from gestures, prosodic
features and body positions and signals respec-
tively (D'Mello, Craig, Witherspoon, McDaniel,
& Graesser, 2008). This information is mapped to
emotional states. The challenge of this approach
Search Nedrilad ::

Custom Search