본문 바로가기
  • Home

A Study on LLM API-Based Mathematical Learning Support System for Visually Impaired Students

  • Journal of Software Forensics
  • Abbr : JSAV
  • 2026, 22(1), pp.41~55
  • Publisher : Korea Software Assessment and Valuation Society
  • Research Area : Engineering > Computer Science
  • Received : March 1, 2026
  • Accepted : March 20, 2026
  • Published : March 31, 2026

Sun-Young Bae 1 Jung, Heo-kyung 1

1배재대학교

Accredited

ABSTRACT

The purpose of this study is to explore the possibility that the LLM API-based interactive explanation and problem generation system can contribute to the improvement of learning ability and learning accessibility of visually impaired students. To this end, a Generative AI-based learning support system was designed and developed to reconstruct mathematical concepts and generate step-by-step example problems according to the context of learners' questions. As a result of application, language-oriented sequential explanations and step-by-step problem-generating structures tended to promote conceptual understanding and improve problem-solving accuracy. In addition, the repetitive confirmation and interaction process strengthened the sense of learning control and self-direction, confirming the possibility of an efficient learning support system.This system can substantially expand accessibility and alleviate the information gap in a digital-based learning environment by structuring the entire learning process beyond simple replacement of information. However, since there is a limitation of small-scale development research, it is necessary to systematically verify the learning effect and policy applicability through long-term and large-scale empirical research in the future.

Citation status

* References for papers published after 2024 are currently being built.