DomainBuilder: the knowledge authoring system for SlideTutor Intelligent Tutoring system

Eugene Tseytlin, Faina Linkov, Melissa Castine, Elizabeth Legowski, Rebecca S. Jacobson
2018 F1000Research  
One of the major challenges in the development of medical Intelligent Tutoring Systems (ITS) is the development of authored content, a time-consuming process that requires participation of discipline experts. In this publication, we describe the development of software systems called DomainBuilder and TutorBuilder, designed to streamline and simplify the authoring process for general medical ITSs. The aim of these systems is to allow physicians without programming or ITSs background to create a
more » ... kground to create a domain knowledge base and author tutor cases in a time efficient manner. DomainBuilder combined knowledge authoring, case authoring, and validation tasks into a single work environment, enabling multiple authoring strategies. Natural Language Processing (NLP) methods were integrated for parsing existing clinical reports to speed case authoring. Similarly, TutorBuilder was designed to allow users to customize all aspects of ITSs, including user interface, pedagogic module, feedback module, etc. Both systems underwent formal usability studies with physicians specializing in dermatology. Open-ended questions assessed usability of the system and satisfaction with its features. Incorporating feedback from usability studies, DomainBuilder and TutorBuilder systems were deployed and used across multiple universities to create customized medical tutoring curriculum. Overall, both systems were well received by medical professionals participating in usability studies with participants highlighting ease of utilization and clarity of presentation. Usability study participants were able to successfully use the system for the authoring tasks. DomainBuilder and TutorBuilder are novel tools that combine comprehensive aspects of content creation, including creation of domain ontologies, case authoring, and validation.
doi:10.12688/f1000research.16060.1 fatcat:j4iwtyj2u5a53bgfi4cig3iwjq