Hung K. Chau

Academic choice and exploration are essential aspects of undergraduate education in the United States, allowing students to select courses with minimal restrictions. However, students often face challenges in navigating the complex academic landscape, hindered by limited information, insufficient guidance, and an overwhelming number of choices. Time constraints from the academic calendar and high demand for popular courses make a thorough evaluation of options difficult. Although academic institutions provide career guidance counselors or advisers, the number of advisers is still limited. Course recommendation systems aim to offer personalized suggestions based on students’ academic backgrounds, preferences, skills, and career goals. However, there is a lack of research on students’ perceptions of recommendations and the provision of explanations to help them evaluate course relevance. Moreover, the majority of course recommender systems concentrate only on the context of learning in higher education. Despite the importance of career goals, none have attempted to establish a connection between learning and work by incorporating job information into course recommendation and explanation. This dissertation explores the development of an advanced course recommendation system in higher education, linking academic courses to career paths using deep learning and natural language processing techniques. It begins by examining various methods for representing and recommending courses, utilizing institutional big data and combining content-based and collaborative models for better performance. A central aspect of this dissertation is the development of skill-based explanations for course recommendations. This is achieved by employing a deep concept extraction model that utilizes BERT and BI-LSTM-CRF architectures. This model effectively extracts concepts from course descriptions, enhancing the recommendation process. Using this concept extraction model, I investigate the impact of skill-based explanations in a serendipitous course recommendation system, which was tested using the AskOski system at the University of California, Berkeley. The findings indicate that these explanations not only increase user interest, particularly in courses with high unexpectedness, but also bolster decision-making confidence. To achieve greater personalization in course recommendations, the future of this field extends beyond academics by incorporating insights from the job market to align with students’ career aspirations. This dissertation introduces an innovative approach for integrating skill-related data into course recommendation systems, effectively bridging the gap between academic pursuits and career aspirations. Finally, I develop an explainable, personalized course recommendation system that incorporates insights from the job market. This system tailors course suggestions based on students’ academic histories and career preferences. Its objective is to enhance course exploration in higher education, assisting students in navigating their educational paths and acquiring the essential skills required for their chosen majors and future careers. A user study conducted at the University of Pittsburgh demonstrates that the recommendations were generally perceived as valuable, with explanations playing a pivotal role in aiding students to assess their interest in the recommended courses. This underscores the significance of integrating skill-related data and explanations into educational recommendation systems.

Higher education is a source of skill acquisition for many middle- and high-skilled jobs. But what specific skills do universities impart on students to prepare them for desirable careers? In this study, we analyze a large novel corpora of over one million syllabi from over eight hundred bachelors’ granting US educational institutions to connect material taught in higher education to the detailed work activities in the US economy as reported by the US Department of Labor. First, we show how differences in taught skills both within and between college majors correspond to earnings differences of recent graduates. Further, we use the co-occurrence of taught skills across all of academia to predict the skills that will be taught in a major moving forward. Our unified information system connecting workplace skills to the skills taught during higher education can improve the workforce development of high-skilled workers, inform educational programs of future trends, and enable employers to quantify the skills of potential workers.

The increasing popularity of digital textbooks as a new learning media has resulted in a growing interest in developing a new generation of adaptive textbooks that can help readers to learn better through adapting to the readers’ learning goals and the current state of knowledge. These adaptive textbooks are most frequently powered by internal knowledge models, which associate a list of unique domain knowledge concepts with each section of the textbook. With this kind of concept-level knowledge representation, a number of intelligent operations could be performed, which include student modeling, adaptive navigation support, and content recommendation. However, manual indexing of each textbook section with concepts is challenging, time-consuming, and prone to errors. Modern research in the area of natural language processing offers an attractive alternative, called automatic keyphrase extraction. While a range of keyphrase and concept extraction methods have been developed over the last twenty years, few of the known approaches were applied and evaluated in a textbook context. In this paper, we present FACE, a supervised feature-based machine learning method for automatic concept extractions from digital textbooks. This method has been created for building domain and student models that form the core of intelligent textbooks. We evaluated FACE on a newly constructed full-scale dataset by assessing how well it approximates concept annotations produced by human experts and how well it supports the needs of student modeling. The results show that FACE outperforms several state-of-the-art keyphrase extraction methods.

Higher education at scale, such as in the California public post-secondary system, has promoted upward socioeconomic mobility by supporting student transfer from 2-year community colleges to 4-year degree granting universities. Among the barriers to transfer is earning enough credit at 2-year institutions that qualify for the transfer credit required by 4-year degree programs. Defining which course at one institution will count as credit for an equivalent course at another institution is called course articulation, and it is an intractable task when attempting to manually articulate every set of courses at every institution with one another. In this paper, we present a methodology towards making tractable this process of defining and maintaining articulations by leveraging the information contained within historic enrollment patterns and course catalog descriptions. We provide a proof-of-concept analysis using data from a 4-year and 2-year institution to predict articulation pairs between them, produced from machine translation models and validated by a set of 65 institutionally pre-established course-to-course articulations. Finally, we create a report of proposed articulations for consumption by the institutions and close with a discussion of limitations and the challenges to adoption.