Yale NLP/LLM Interest Group - Session 11 Xueqing Peng

Speaker Bio:

Xueqing Peng, PhD, is a Postdoctoral Associate at the Section of Biomedical Informatics and Data Science, Yale School of Medicine. Previously, she was a Postdoctoral Research Fellow at the School of Biomedical Informatics, University of Texas Health Science Center at Houston. Xueqing graduated with a doctoral degree in Medical Systems Biology from Fudan University in 2022. Xueqing’s research interests include medical data science, machine learning, and natural language processing (NLP).

Title of Talk: A New Approach to Detecting Semantic Novelty of Biomedical Literature

Identifying novel research work is a significant challenge in the rapidly evolving field of biomedical research. We propose a novel approach to assess the semantic novelty of scientific publications by leveraging recent LLM-based text embedding models. Using the text of publications, such as titles and abstracts, we build a "semantic universe" that maps the landscape of biomedical scientific knowledge, letting us measure the semantic novelty of each article based on its text embedding. To evaluate our approach, we examined its correlations with other measurements, such as citation count and other relevant metrics. Additionally, we explored its potential applications in visualizing the evolution of research fields overtime and recognizing breakthrough research, such as Nobel Prize-winning work. We believe that our approach can serve as a valuable tool to assist researchers in exploring new areas, provide deeper insights to funding agencies when considering proposals, and offer publishers an additional perspective in evaluating research.