The Centre for Analytics and Artificial Intelligence Engineering (CARTE) has created an in-development artificial intelligence (AI) tool named Education Pathways to help U of T students with course selection. The tool aims to provide students with an alternative to the Course Finder website by allowing users to match search terms with relevant courses, along with recommending early-year courses that will best prepare students for courses they may take in their following years of study.
The tool accepts a number of filters, including year of study, faculty, and campus, along with a search term. It then offers a list of courses that include not only the search term, but also related terms, as well as potential prerequisites for other searched courses.
“A more intelligent search engine”
This tool was developed after a series of focus group meetings in September with students from the Faculty of Applied Science & Engineering (FASE) to get feedback on the expectations and experiences students had in analytics, AI, and machine learning (ML) education. In response to this feedback from students, the CARTE worked on a series of initiatives, including the AI, to improve student experiences.
In an email to The Varsity, the CARTE’s assistant director, Somayeh Sadat, wrote that students felt there were not enough advanced, niche, or skill-oriented courses in the analytics, AI, and ML course offerings, though these types of courses were, in fact, offered by the university. They found that the existing course finder was not helpful in finding these courses, as it could only search based on the exact terms put in by the user.
“More importantly, not all courses had listed prerequisites and students needed guidance on how to best prepare for advanced courses,” Sadat added.
In response, Education Pathways has been one of the recent initiatives generated to assist students. The tool works to offer guidance on courses from any area of study, including analytics, AI, ML, humanities, arts, and sciences.
Sadat considered Education Pathways to be “a more intelligent search engine compared to the course finder” for various reasons. One reason is that the tool can search all relevant courses more flexibly by exploring both the terms entered directly by the user and any other related terms. Another reason is that it suggests prerequisite courses to aid students in deciding what courses they will take in later years.
Developing the tool
According to Sadat, the process of developing Education Pathways began with the CARTE’s research assistant, Alex Olson, using the information found in Course Finder to map out every term used in the course descriptions across U of T. Based on the information this gathered, Olson wrote an algorithm.
Sadat further explained that “by looking at the relationships between different words, it was then possible to predict the relevance of a course to any search term, even if the search only uses words that don’t come up in the course’s actual description at all.”
Finally, academic staff, students, and FASE leadership provided feedback about the tool, and in early 2021, Education Pathways was launched.
Members of CARTE have been collecting feedback and are still asking for suggestions and comments from students on Education Pathways. With this feedback, CARTE can assess any future areas of improvement and the level of success of the new course search tool.
Sadat described possible improvements that could be implemented to enhance the tool and help with student pathways.
“There is potential to build on this work and supplement it with student registration data in order to better understand student pathways through electives and interdivisional connections and showcase to external partners, as well as identify opportunities to build new programs,” Sadat wrote.
According to Sadat, CARTE sees high potential in improving Education Pathways so that it can better serve students’ course planning and assist them in gaining a better understanding of the latest courses.