HEAL Projects

Structuring the Knowledge Landscape of Dental Informatics through AI-driven Keyword Extraction and Taxonomy Design

This project evaluates automated keyword extraction and taxonomy generation methods for dental informatics literature published from 2018 to 2023. Using a curated dataset of 1,252 articles from PubMed, we compared YAKE and ChatGPT 3.5 against author-supplied keywords, with ChatGPT 3.5 demonstrating superior precision, recall, and broader coverage. We further developed and evaluated both bottom-up (BERT-based clustering) and top-down (ChatGPT-prompted) taxonomies, assessed by human experts for correctness and completeness. Our findings reveal a highly concentrated authorship pattern and dispersed journal landscape, highlighting the fragmented nature of the field. This work advances scalable techniques for literature organization, supports research trend analysis, enhances information retrieval, and lays the groundwork for future applications in dental education and clinical decision support.

Disaster Informatics after the COVID-19 Pandemic

This study presents a comprehensive bibliometric and topic analysis of the disaster informatics literature published between January 2020 to September 2022. Leveraging a large-scale corpus and advanced techniques such as pre-trained language models and generative AI, we identify the most active countries, institutions, authors, collaboration networks, emergent topics, patterns among the most significant topics, and shifts in research priorities spurred by the COVID-19 pandemic. Our findings highlight (1) countries that were most impacted by the COVID-19 pandemic were also among the most active, with each country having specific research interests, (2) countries and institutions within the same region or share a common language tend to collaborate, (3) top active authors tend to form close partnerships with one or two key partners, (4) authors typically specialized in one or two specific topics, while institutions had more diverse interests across several topics, and (5) the COVID-19 pandemic has influenced research priorities in disaster informatics, placing greater emphasis on public health. We further demonstrate that the field is converging on multidimensional resilience strategies and cross-sectoral data-sharing collaborations or projects, reflecting a heightened awareness of global vulnerability and interdependency. Collecting and quality assurance strategies, data analytic practices, LLM-based topic extraction and summarization approaches, and result visualization tools can be applied to comparable datasets or solve similar analytic problems. By mapping out the trends in disaster informatics, our analysis offers strategic insights for policymakers, practitioners, and scholars aiming to enhance disaster informatics capacities in an increasingly uncertain and complex risk landscape.

Related papers: Tran, N., Chen, H., Cleveland, A., & Zhou, Y. Disaster Informatics after COVID-19 Pandemic: Bibliometrics and Topic Analysis. Under review.

Innovation Detection and Analysis for Interdisciplinary Research

This project seeks to harness the latest advances in LLMs to transform how groundbreaking biomedical research is proposed and funded. Biomedical science relies heavily on rigorous grant proposals to drive research into pressing areas such as drug discovery, disease mechanisms, and therapeutic innovation. By assembling a comprehensive dataset of existing grant proposals, scientific publications, patents, and clinical trial records, this project systematically evaluates the capacity of LLMs to produce high-quality, innovative, and feasible proposal ideas specifically tailored to the biomedical domain. The research includes thorough experimentation to gauge the originality, reliability, and practical viability of AI-generated suggestions—factors that are crucial for funding bodies assessing new projects in complex areas like genomics, personalized medicine, and regenerative therapies. Ultimately, these findings will provide critical insights into how AI can streamline and enrich the scientific funding landscape, offering a powerful tool for researchers, funding agencies, and the broader biomedical community to accelerate the pace of life-saving discoveries.

Related papers: Wang, Z., Wang, N., Zhang, H., Wang, Z., Wang, Z., Ding, J., & Chen, H. (2025). IBID-CCT: A novel model for interdisciplinary breakthrough innovation detection based on the cusp catastrophe theory. Information Processing & Management, 62(4), 104121.

Data Quality for Medical Concept Normalization

Medical concept normalization (MCN) aims to map informal medical terms to formal medical concepts, a critical task in building machine learning systems for medical applications. However, most existing studies on MCN primarily focus on models and algorithms, often overlooking the vital role of data quality. This research evaluates MCN performance across varying data quality scenarios and investigates how to leverage these evaluation results to enhance data quality, ultimately improving MCN performance through the use of large language models (LLMs). The effectiveness of the proposed approach is demonstrated through a case study. We begin by conducting a data quality evaluation of a dataset used for MCN. Based on these findings, we employ ChatGPT-based zero-shot prompting for data augmentation. The quality of the generated data is then assessed across the dimensions of correctness and comprehensiveness. A series of experiments is performed to analyze the impact of data quality on MCN model performance. These results guide us in implementing LLM-based few-shot prompting to further enhance data quality and improve model performance.

Related papers:

Chen, H., Chen, J., & Ding, J. (2021). Data evaluation and enhancement for quality improvement of machine learning. IEEE Transactions on Reliability, 70(2), 831-847.

Chen, H., Li, R., Cleveland, A., Ding, J.(2025). Enhancing data quality in medical concept normalization through large language models..Journal of Biomedical Informatics, 104812.

Evaluating Generative AI for Domain-specific Applications

With the advancement of AI-powered language models, such as ChatGPT, generative AI (AIGC, a.k.a AI-generated content) are being discussed and used in every aspect of human society. Generative AI has strong ability to analyze and create text, images, code, and beyond. In addition to the fundamental techniques behind AIGC and the popular tasks of AIGC, it is also being widely applied to different fields and domains, including question answering, marketing, healthcare, gaming, music, drug discovery, etc. Researchers and scientists have also started discussing the general design principles for generative AI applications. However, it is undeniable that the generative AI revolution is still in the early stage, and it can change and improve with the potential for even greater future capability. This project aims to investigate: (1) Frameworks and methodologies of evaluating the capacity and limitations of Generative AI in specific domains. (2) The applications of Generative AI in high stake domains, such as legal, medical, cyber security, and others. (3) Generative AI for information extraction, information retrieval, questions answering, text summarization, and other similar tasks. (4) Data argumentation with Generative AI. (5) Quality assurance for Generative AI. (6) Methodologies of quality evaluation of AI generated content. (7) Generative AI for future teaching and learning.

Related papers: Ding, J., Chen, H., Kolapudi, S., Pobbathi, L., & Nguyen, H. (2023, October). Quality Evaluation of Summarization Models for Patent Documents. In 2023 IEEE 23rd International Conference on Software Quality, Reliability, and Security (QRS) (pp. 250-259). IEEE.