aalto1 untyped-item.component.html
Stratux: A user-centered design of multi-agent learning system
Loading...
URL
Journal Title
Journal ISSN
Volume Title
School of Science |
Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.
Authors
Date
Department
Major/Subject
Mcode
Language
en
Pages
77
Series
Abstract
Large language models (LLMs) have created new opportunities for self-directed learning, enabling students with foundational skills to engage in personalized, ondemand study. In this work, I leverage recent advances in LLMs to scaffold structured, self-directed learning environment. Consequently, I developed Stratux, a multi-agent conversational learning system. It features pedagogical agents that allow learners to collaboratively explore learning goals, while integrating a Bloom’s taxonomy-based assessment module to help learners evaluate their progress. This thesis details the user-centered design process and describes how prompt engineering was utilized to tailor LLM interactions for educational contexts. A user study with 18 students aged 16–25 shows that interacting with Stratux leads to a lower subjective workload compared to using a general-purpose LLM. However, the study also reveals variations in learner engagement and performance across different learning topics and cognitive levels. In conclusion, I argue that multi-agent LLM applications like Stratux can offer a unique, structured, and comprehensive learning experience for learners seeking to guide their own education.