"TISBI" on DAMDID: How LLMs Help Find Similar Graduate Pathways and Suggest Bachelor's and Master's Theses Topics

The conference brought together researchers, lecturers, students, postgraduates, and practising specialists in the field of artificial intelligence, data analysis, and management. On 21 October, Ramil Khabibrakhmanov presented the work "«Student Representation Methods for Alumni Retrieval and LLM-Based Thesis Topic Recommendation»".
 
 The study is dedicated to the construction and comparison of methods for representing a student's academic profile. The quality of such representations was assessed in the task of finding graduates with similar educational pathways and subsequent generation of final qualifying work topics using a language model.

"I am glad that we were able to present our research at such a conference." It was not only the presentation itself that was valuable, but also the work on the article during the peer review stage – an excellent opportunity to receive feedback from experts in the field of Data Science, learn about alternative methodological approaches, and adopt some of their experience. The conference itself is, of course, a multitude of interesting ideas and people. I remember the data analysis in education section the most. It was interesting to learn what tasks other researchers set and what approaches they use to solve them, and then to compare their work with ours"Ramil Khabibrakhmanov