Applied machine learning in intelligent systems - knowledge graph-enhanced ophthalmic contrastive learning with “clinical profile” prompts

Aug 5, 2026

Upload by Riley

Artificial intelligence can improve ophthalmic diagnosis accuracy and interpretability, yet limited transparency reduces clinician trust in automated recommendations. This study integrates knowledge graphs with contrastive learning and “clinical profile” prompts to refine the ophthalmology-oriented large language model MeEYE based on the CHATGLM3-6B architecture. Fine-tuning with structured clinical knowledge enhances feature discrimination, while the prompts improve contextual understanding. Comprehensive benchmarking and clinical case studies validate gains in diagnostic accuracy and interpretability through transparent, clinically relevant AI recommendations.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Introduction

Methods

Results

Ethics, Funding, and Data Availability

TYPE Original Research PUBLISHED 12 March 2025 DOI 10.3389/frai.2025.1527010 OPEN ACCESS EDITED BY Olawande Daramola, University of Pretoria, South Africa REVIEWED BY Nikos Kanakaris, University of Southern California, United States Nan Chen, Eye Hospital of Nanjing Medical University, China *CORRESPONDENCE Mini Han Wang [d20092100037@cityu. mo](d20092100037@cityu. mo)[ ](d20092100037@cityu. mo)Guanghui Hou [Houguanghui901@163.com](Houguanghui901@163.com)[ ](Houguanghui901@163.com)Junbin Fang [tjunbinfang@jnu.edu.cn](tjunbinfang@jnu.edu.cn)[ ](tjunbinfang@jnu.edu.cn)Xiangrong Yu [yxr00125040@126.com](yxr00125040@126.com)[ ](yxr00125040@126.com)Kelvin Kam-Lung Chong [chongkamlung@cuhk.edu.hk](chongkamlung@cuhk.edu.hk)[ ](chongkamlung@cuhk.edu.hk)Yi Pan [yi.pan@siat.ac.cn](yi.pan@siat.ac.cn)[ ](yi.pan@siat.ac.cn)RECEIVED 12 November 2024 ACCEPTED 26 February 2025 PUBLISHED 12 March 2025 CITATION Han Wang M, Cui J, Lee SM-Y, Lin Z, Zeng P, Li X, Liu H, Liu Y, Xu Y, Wang Y, Alves JLCDC, Hou G, Fang J, Yu X, Chong KK-L and Pan Y (2025) Applied machine learning in intelligent systems: knowledge graph-enhanced ophthalmic contrastive learning with “clinical profile” prompts. Front. Artif. Intell. 8:1527010 . doi: 10.3389/frai.2025.1527010 COPYRIGHT © 2025 Han Wang, Cui, Lee, Lin, Zeng, Li, Liu, Liu, Xu, Wang, Alves, Hou, Fang, Yu, Chong and Pan. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Applied machine learning in intelligent systems: knowledge graph-enhanced ophthalmic contrastive learning with “clinical profile” prompts Mini Han Wang 1,2*, Jiazheng Cui3,4, Simon Ming-Yuen Lee 5, Zhiyuan Lin 6, Peijin Zeng 6,7, Xinyue Li 8, Haoyang Liu 9,10, Yunxiao Liu 9,10, Yang Xu7, Yapeng Wang 9, José Lopes Camilo Da Costa Alves 11, Guanghui Hou 12*, Junbin Fang 13*, Xiangrong Yu 14*, Kelvin Kam-Lung Chong 2* and Yi Pan 15* 1Zhuhai Precision Medical Center, Zhuhai People's Hospital, The Affiliated Hospital of Beijing Institute of Technology, Zhuhai Clinical Medical College of Jinan University, Zhuhai, China, 2 Department of Ophthalmology and Visual Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong SAR, China, 3Zhuhai Institute of Advanced Technology, Chinese Academy of Sciences (CAS), Zhuhai, China, 4 Beijing Normal University-Hong Kong Baptist University United International College, Zhuhai, China, 5 Department of Food Science and Nutrition, Hong Kong Polytechnic University, Kowloon, Hong Kong SAR, China, 6 Perspective Technology Group, Zhuhai, China, 7 Beijing Institute of Technology, Zhuhai, China, 8 Department of Ophthalmology, Tianjin Medical University, Tianjin, China, 9 Faculty of Applied Sciences, Macao Polytechnic University, Macao, Macao SAR, China, 10 Digital Healthcare and Artificial Intelligence Association, Macao, Macao SAR, China, 11 Faculty of Business, City University of Macau, Macao, Macao SAR, China, 12Zhuhai Aier Eye Hospital, Zhuhai, China, 13College of Science & Engineering, Jinan University, Shenzhen, China, 14Zhuhai People's Hospital (Zhuhai Clinical Medical College of Jinan University), Zhuhai, China, 15Shenzhen Key Laboratory of Intelligent Bioinformatics, Shenzhen Institute of Advanced Technology, Shenzhen, China Introduction: The integration of artificial intelligence (AI) into ophthalmic diagnostics has the potential to significantly enhance diagnostic accuracy and interpretability, thereby supporting clinical decision-making. However, a major challenge in AI-driven medical applications is the lack of transparency, which

FAQs

What problem does the study address in AI-driven ophthalmic diagnostics?

The study targets limited transparency in AI recommendations, which weakens clinician trust and restricts clinical decision-making.

How does the method improve the MeEYE ophthalmology model?

It combines knowledge graph-based domain knowledge with contrastive learning and adds “clinical profile” prompts to strengthen contextual understanding and clinically relevant feature capture.

What outcomes are reported in the results?

Integrating knowledge graphs and contrastive learning improves diagnostic accuracy and model interpretability, enabling more precise and clearer identification of ophthalmic conditions and more transparent recommendations.

Aug 5, 2026

Upload by Riley