Assessing the Proficiency of Large Language Models on Funduscopic Disease Knowledge (Preprint)

Author:

Shao YiORCID,Wu Jun-YiORCID,Zeng Yan-Mei,Qian Xian-Zhe,Hong Qi,Hu Jin-Yu,Wei Hong,Zou Jie,Chen Cheng,Wang Xiao-Yu,Chen Xu

Abstract

BACKGROUND

Large language models (LLMs) have significantly transformed the field of natural language processing, with cutting-edge models like ChatGPT currently leading the way in medical AI.

OBJECTIVE

This study aimed to assess the performance of five distinct LLMs (GPT-3.5, ChatGPT-4, PaLM2, Claude 2, and SenseNova) in comparison to two human cohorts (a group of funduscopic disease experts and a group of ophthalmologists) on the specialized subject of funduscopic disease.

METHODS

Five distinct LLMs and two distinct human groups independently completed a 100-item funduscopic disease test. The performance of these entities was assessed by comparing their average scores, response stability, and answer confidence, thereby establishing a basis for evaluation.

RESULTS

Among all the LLMs, GPT-4 and PaLM2 exhibited the most substantial average correlation. Additionally, GPT-4 achieved the highest average score and demonstrated the utmost confidence during the exam. In comparison to human cohorts, GPT-4 exhibited comparable performance to ophthalmologists, albeit falling short of the expertise demonstrated by funduscopic disease specialists.

CONCLUSIONS

The study provided evidence of the exceptional performance of GPT-4 in the domain of funduscopic disease. With continued enhancements, validated LLMs have the potential to yield unforeseen advantages in enhancing healthcare for both patients and physicians.

Publisher

JMIR Publications Inc.

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3