We are looking for a postdoctoral fellow to work on a new research project aimed at creating a comprehensive platform for authoring and running adaptive online courses. The goal of the authoring environment is to personalize instruction (e.g., with recommendations or other forms of guidance) according to learners' prior knowledge, knowledge growth, interests, and other factors. Our initial strategy will be to integrate CMU's Open Learning Initiative (OLI, an online course platform) and CMU's Cognitive Tutor Authoring Tools (CTAT, a set of authoring tools for creating adaptive tutoring systems). Course authors will have a variety of ways to personalize instruction in online courses: we envision a student model and adaptivity in all of the system's regulative loops. An additional goal is to support AB testing to enable routine tests of the benefits to learners of different forms of adaptivity.
This work will be a step toward a broad framework for adaptive services for technology-enhanced learning. To enable integration and interoperability with other platforms and tools, the system will adhere to emerging IEEE Adaptive Instructional System Standards and other applicable standards (e.g., xAPI and/or LTI, possibly to include Caliper). Results will be open-source.
The postdoc will be involved in all aspects of the research. Specific responsibilities will be to create a fully browser-based version of CTAT and to achieve, in stages, a progressively more complete technical integration of CTAT and OLI. We plan to integrate both for student run time, where CTAT tutors are embedded in OLI courses, and for course authoring time, to permit changes to CTAT tutors from OLI's authoring environment. The integration will entail information exchange between the systems' student models and dashboards and will permit authors to plug in alternative adaptive algorithms. We want to extend current conceptualizations of adaptivity typically found in intelligent tutoring systems, like adaptive problem selection, in order to apply them to the content of online courses, where the system can make adaptive choices from a wider selection of materials. We aim to create realistic demonstrations of these new kinds of adaptivity by transforming existing (currently not-so-adaptive) online courses. We expect that the postdoc would then want to explore any of a whole host of research questions that this kind of system presents. For example, what sort of algorithms are needed to choose properly among disparate types of online materials? Does the system increase authoring ease and efficiency? Are the resulting courses easy enough to maintain, share and customize as to facilitate adoption in the real world?
PhD in learning technologies, human-computer interaction, or a related field
Experience with web development
Skill in interaction design and usability testing
Experience with complex software engineering projects
Experience with e-learning architecture design or development
Experience designing and developing authoring tools, for learning technologies or other technologies
Experience design and running controlled experiments (AB tests) that involve evaluations of learning
Experience with Vue.js or a similar front-end framework