摘要
Adaptive learning solutions generally assume curricular, instructional, content, and/or learning models to be known in advance. This paper describes an ongoing research project being conducted by the Center for Inclusive Software for Learning (CISL), funded by the U.S. Department of Education, that is developing a suite of open-source tools designed for diverse learners by making digital educational materials—including open educational resources (OER)—accessible, flexible, and engaging. The Clusive Reader at the heart of CISL leverages Universal Design for Learning (UDL) principles and includes a number of adaptive features designed to operate without prior knowledge of how the materials support learning goals. These adaptive features include content complexity leveling, inclusion of summaries and highlighted main ideas, dictionary and glossary tools, and just-in-time tips and suggestions regarding preference setting and use of learning tools. Information to support adaptivity heuristics comes from user interactions, an onboarding skills discovery process, and embedded affective, comprehension, and learning goal prompts and game-like vocabulary challenges. Per UDL, adaptivity in Clusive is designed to support not only immediate learning goals, but also the building of “expert learners” who independently choose the most effective content, contexts, and supports. As such, adaptivity has been designed as a learning scaffold, with the intention of gradual reduction over time, in two ways. First, adaptivity operates transparently, providing students with insights as to why choices and recommendations are made. Second, students are allowed to override system decisions to promote self-agency. Evaluation of student impressions and efficacy are being conducted as part of the CISL project. This paper provides the research foundations and philosophical approach taken for this effort, describes the current adaptive learning features being designed, and concludes with remarks on next steps and challenges being faced.