
Research Article
Building a Chatbot to Adopt an Effective Learning Strategy for Graduate Courses in Computer Science
@INPROCEEDINGS{10.1007/978-3-031-84312-9_19, author={Alex Elentukh and Shravan Motipally and Jacob Kustra and Toby March}, title={Building a Chatbot to Adopt an Effective Learning Strategy for Graduate Courses in Computer Science}, proceedings={Computer Science and Education in Computer Science. 20th EAI International Conference, CSECS 2024, Sofia, Bulgaria, June 28--30, 2024, Proceedings}, proceedings_a={CSECS}, year={2025}, month={3}, keywords={computer science education software engineering teaching and learning effectiveness language models chatbots prompt engineering retrieval augmented generation}, doi={10.1007/978-3-031-84312-9_19} }
- Alex Elentukh
Shravan Motipally
Jacob Kustra
Toby March
Year: 2025
Building a Chatbot to Adopt an Effective Learning Strategy for Graduate Courses in Computer Science
CSECS
Springer
DOI: 10.1007/978-3-031-84312-9_19
Abstract
Courses offered in a Computer Science program of a large university differ greatly in their structure. Such a variety is driven by the diversity of backgrounds of faculty members, as well as by the breadth and depth of topics covered. The pragmatic goal of our research is to help students to adopt an effective learning strategy while engaging in a graduate program of several dozen very different courses. As a first step, we established a database of frequently asked questions regarding class logistics. Building on this foundation we launched the chatbot to be readily available to all program registrants. This paper can serve as a step-by-step introduction to How-To-Build-A-Chatbot while dealing with challenges and successes encountered. A student is able to quickly browse through relevant questions and responses. Additionally, a student can actively participate in a conversation with the bot, select among five language models, adjust semantic matching, and switch on generative mode. We used Ragas framework to formally evaluate and measure the performance of the bot, while focusing on its ability to simulate the human-like interaction. The experimental nature of exchange with the bot served as a strong motivation for students to keep exploring and learning. The evolution from a list-bound bot design into a more fluid approach involving language technologies is not trivial. As a benefit, we were able to offer students the personalized and engaging learning experience.