Efficacy of large language models in facilitating exam preparation for medical students

Abstract

The complex medical curriculum compels the students to study a significant quantity of textbooks authored by various writers in each subject. These books contain a considerable amount of topics, and as a consequence, students frequently become overwhelmed by the heavy load of academic pressure. Moreover, in the four phases of medical studies, students perform six cards, three terms, and one professional examination in each phase to advance to the next phase. In this particular circumstance, students need adequate guidance to clarify their understanding and to address any questions that may arise during their preparation. To resolve these issues, an expert LLM developed with RAG, especially for medical students, is required. We intend to develop a retriever model powered by the proficiency of LLMs that will have access to the comprehensive resources of the medical curriculum and is able to solve any study-related confusion a medical student may have. We believe that our algorithm will help medical students overcome their fear of examinations, as well as improve their overall academic performance. LLM has great potential to give the perfect guide a medical student requires. LLMs integrated with the RAG system has the ability to give a medical student the perfect tutoring required to test their exam preparation. Our driving force motivation behind implementing this algorithm is to lessen the mental pressure of medical students and enhance their study experience by providing curriculum-aligned assessment tools.