Collaborative Research: Open Player and Community Modeling as a Learning Tool


PIs: Magy Seif El-Nasr, Northeastern University (Award Details)
Jichen Zhu, Drexel University (Award Details)

An active research area in intelligent tutoring systems and game-based learning is personalized learning. In these environments, the computer analyzes learner behaviors in real time and builds individual models of critical aspects of learning (e.g., level of engagement, knowledge acquisition) to adapt the system. While these player models capture essential elements of a learner, the valuable information they contain is usually opaque and hidden from the learners. As a result, players of adaptive learning games do not know how the game categorizes them and why the game changes. This then affects their learning, especially when learning complex concepts. To address this, the project will investigate a novel approach to personalized learning environments by providing learners with a toolset, called Open Player and Community Model, which allows learners to see their actions, tactics, and strategies as well as those of others. A playback system with a labeling toolset will also be created to allow teachers and other learners to annotate patterns they see others use, which will shed light on strategies and mental models of solutions, errors, corrections, and misconceptions.

The project will explore how students learn problem-solving with tools that allow them to reflect on their practices and on how others solved similar problems. The research will study two questions: First, how do people engage with user/player models when they are made transparent and used to reflect on learning? Second, how can designers create meaningful visualizations for learners and instructors to support reflection on user/player models and crowdsourced annotations of problem solving behavior? Tools to support collecting and visualizing open player and community models will be developed and made available to the public. Studies of the tools will be conducted to determine how they support learning in complex domains.

Tags: , ,