Related Experiment Video
Updated: Jun 7, 2025
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
504
Large Language Models in Dental Licensing Examinations: Systematic Review and Meta-Analysis
Mingxin Liu1, Tsuyoshi Okuhara2, Wenbo Huang3
1Department of Health Communication, Graduate School of Medicine, The University of Tokyo, Bunkyo, Tokyo, Japan.
International Dental Journal
|November 12, 2024
Summary
Large language models (LLMs) show promise for dental education, but current accuracy in licensing exams is insufficient for clinical use. GPT-4 performed best, though more dental-specific training data is needed.
Area of Science:
- Artificial Intelligence in Dentistry
- Medical Education Technology
Background:
- Large language models (LLMs) are increasingly integrated into various professional fields.
- Their application in specialized domains like dentistry requires rigorous evaluation.
Purpose of the Study:
- To systematically review and meta-analyze the performance of LLMs in global dental licensing examinations.
- To assess LLM accuracy across different linguistic and geographical contexts.
- To inform potential applications in dental education and diagnostics.
Main Methods:
- Systematic review and meta-analysis following PRISMA guidelines.
- Searches conducted on PubMed, Web of Science, and Scopus (Jan 2022 - May 2024).
- Quality assessment using QUADAS-2, with qualitative and quantitative performance analyses.
Main Results:
- Eleven studies from eight countries were included.
- GPT-4 achieved 72% accuracy, outperforming GPT-3.5 (54%) and Bard (56%).
- GPT-3.5 showed better performance in English-speaking regions; GPT-4's performance was consistent globally.
Conclusions:
- LLMs, especially GPT-4, demonstrate potential but lack the accuracy for current clinical dental applications.
- Insufficient dental training data and challenges with image-based diagnostics limit LLM performance.
- Current LLMs are not yet suitable for dental education or clinical diagnosis.
