The employment landscape is changing and expanding, creating uncertainty as individuals navigate decisions regarding career preparation and continuing education. This project develops technology to provide guidance for college students to access high quality information and guidance within a large and complex decision space. The approach includes integrating social support from human career coaches or peers to increase confidence and motivation in the problem solving and decision making process. The approach is fundamentally data driven, built upon computational modeling while informed by theories of human behavior, embodied within a novel machine learning paradigm termed Socially-Sensitive Reinforcement Learning (SSRL). Based on this new paradigm, the system will generate guidance to support students while being sensitive to student needs and preferences so that there is a high probability that students will accept and benefit from the guidance. The project targets nontraditional students who embark upon or advance STEM career paths and aims to broaden participation in STEM.
The project contributes to both human-computer interaction and machine learning and takes an interdisciplinary approach in its modeling of human interaction and sociotechnical support for learning. It comprises a three pronged solution: (1) a computational modeling strand proposes new computational paradigms that continually map best practices and inform data-driven recommendations in support of effective decision making; (2) a behavioral research strand provides insight through qualitative and quantitative investigations, uncovering properties of recommendations as well as recommenders that are associated with whether guidance is accepted and how it influences success in career advancement; and (3) an intelligent coach development strand embodies findings from the other two strands in the CareerScope Intelligent Coach Agent (ICA), designed as a sociotechnical solution leading to impact student decision making. The partnership with Western Governor’s University (WGU) facilitates transition of research into practice at large scale through research with WGU students and deployment on the WGU platform. While the research will be housed within a single platform, the technical innovations will ultimately be integrated into a wide range of platforms. The results of this project will broaden understanding of the limitations and opportunities for ICAs to guide learners, particularly non-traditional learners, through their career paths. These results can inform both online and brick-and-mortar universities how to best coach/train/mentor students or integrate peer/professional mentorship and peer interaction into current advising practices.