Modeling Self-Regulated Learning With Intelligent Multiagent Learning Technologies: Beyond Cognition and Metacognition

• In Event: 31.086 - Cognitive, Social, and Motivational Processes Involving Modeling: A Means of Effective Self-Regulatory Instruction and Learning

Abstract

In this presentation, we argue for the integration of cognitive, metacognitive, and affective processes in modeling human-artificial agent interaction during complex learning with intelligent, advanced learning technologies (Azevedo et al., 2013). Several approaches have been used to model learners’ cognitive and metacognitive strategies (e.g., Azevedo & Cromley, 2004). However, each of these approaches has neglected the key role of affect during human-artificial agent (e.g., avatars, intelligent virtual humans) interactions during the modeling of SRL processes. Affective processes need to be incorporated as part of successful modeling as learners typically experience negative affective reactions (e.g., anger, frustration, contempt) when attempting to acquire, internalize, and practice cognitive and metacognitive SRL processes. As such, we argue that the modeling of cognitive and metacognitive processes for human learners must also include artificial agents’ ability to detect, track, and externally regulate human learners’ affective responses.

We will present the results of several recent experimental studies that have examined the effectiveness of using artificial pedagogical agents to support and foster college students’ cognitive and metacognitive self-regulatory skills during complex learning of human biology. Being able to accurately monitor and regulate one’s own cognitive and metacognitive processes and strategies requires a tremendous amount of effort, but is required in order to be a successful learner. Results to be presented will include learning outcomes, self-report measures, and processes data. More specifically, we will elaborate and provide evidence of the integral role of affect in facilitating the understanding, acquisition, integration, and use of cognitive and metacognitive processes during complex learning.

In this presentation, we will focus on defining and illustrating cognitive and metacognitive modeling strategies (Winne & Azevedo, in press); reviewing the empirical literature on multimedia learning of metacognitive strategies as they relate to modeling (Azevedo, in press); providing suggestions for augmenting contemporary cognitive theories of multimedia learning to include emotions (D’Mello et al., 2013; Mayer, in press; Pekrun et al., 2011; Scherer, 2009); and proposing empirically-based principles for designing intelligent, multi-agent learning technology environments aimed at fostering metacognitive strategies to facilitate modeling (Biswas et al., 2010). We will conclude by recommending several areas for future research on modeling SRL with intelligent multi-agent learning technologies.

Authors

• Roger Azevedo, North Carolina State University

• Michelle Taub, North Carolina State University

• Nicholas Vincent Mudrick, North Carolina State University