Magazine article AI Magazine

Reports of the AAAI 2010 Fall Symposia

Magazine article AI Magazine

Reports of the AAAI 2010 Fall Symposia

Article excerpt

Cognitive and Metacognitive Educational Systems

The Cognitive and Metacognitive Educational Systems (MCES) AAAI symposium, held in November 2010, was the second edition of this successful AAAI symposium. The idea for the symposium stemmed from several theoretical, conceptual, empirical, and applied considerations about the role of metacognition and self-regulation when learning with computer-based learning environments (CBLEs). A related goal was the design and implementation issues associated with metacognitive educational systems. MCES implemented as CBLEs are designed to interact with users and support their learning and decision-making processes. A critical component of good decision making is self-regulation.

The primary aim of this symposium was to continue the discussion started in 2009 on some of the previous considerations and to enhance the discussions with some new ones: What are the theoretical foundations and how are they articulated in CBLEs? Is it possible to develop a unified framework for all metacognitive educational systems? What are the necessary characteristics of these systems to support metacognition? To what extent does the educational system itself have to exhibit metacognitive behaviors, and how are these behaviors organized and enacted to support learning? What are the main aspects of metacognition, self-regulation skills, emotions, and motivations that influence the learning process? What does it mean to be metacognitive, and how can one learn to be metacognitive? Can MCES actually foster learners to be self-regulating agents? How can an MCES be autonomous and increase its knowledge to match the learners' evolving skills and knowledge? What is the role of artificial agents in supporting metacognition and self-regulated learning? MCES may not be embodied, but does it help if they act as intentional agents?

This symposium aimed to provide a comprehensive definition of metacognitive educational systems that is inclusive of the theoretical, architectural, and educational aspects of this field. To meet these goals, we stimulated the debate with two panel sessions. The first, chaired by Janet Kolodner (National Science Foundation) explored the questions of what metacognition is, what pieces of it are needed for learning, what pieces need to be learned, and how can it be supported with technology. The second panel, chaired by Roger Azevedo (McGill University) explored measurement issues in SRL.

The symposium hosted many contributions from researchers in heterogeneous disciplines: AI, cognitive and learning sciences, education psychology, education science, human-computer interaction (HCI), computational linguistics, web technologies, social network analysis, visualization techniques, software architectures, and multiagent systems. Discussion focused mainly on the need to have quantitative measures of the learner's metacognitive abilities. The debate was between education psychologists and AI and HCI people. The former need to have measures of metacognition in support of the evidence of particular behaviors in the learner when he or she is engaged in studying a topic. The latter want to have computable models of metacognitive abilities to build a new generation of truly metacognitive agents that are able to support the learning process. Both kinds of people argued that suitable computable models are needed to represent metacognitive processes despite the particular research goal.

The discussion was enriched by three keynote speeches. Kenneth R. Koedinger (Human Computer Interaction Institute, Carnegie Mellon University) discussed using data to make educational systems more metacognitive. Chris Quintana (School of Education, University of Michigan) gave a speech titled "Making the Implicit Explicit: Issues and Approaches for Scaffolding Metacognitive Activity." Gautam Biswas (Center for Intelligent Systems, Vanderbilt University) presented "Modeling and Measuring Self-Regulated Learning in Teachable Agent Environments. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.