报告人：严善楷 副教授 海南大学
报告时间：2024年12月13日（周五）上午 11:00
报告地点：#腾讯会议：176-852-375
报告摘要：Recent advancements in natural language processing have revolutionized the way we analyze and generate biomedical texts. However, purely data-driven models often struggle with domain-specific knowledge gaps, limiting their effectiveness in complex fields like biomedicine. To address this, retrieval-augmented generation (RAG) techniques, powered by knowledge graphs, offer a promising solution by incorporating structured domain knowledge into language models. This talk will explore the integration of knowledge graphs into NLP workflows for biomedicine, focusing on how retrieval-augmented generation enhances the precision and relevance of language understanding and generation. We will discuss key components, including the construction of domain-specific knowledge graphs, their role in augmenting model outputs, and real-world applications such as clinical decision support, medical Q&A, and biomedical literature retrieval. You will gain insights into the technical architecture behind RAG models, the challenges of applying these systems to biomedical data, and strategies to overcome those challenges. By bridging the gap between structured knowledge and unstructured text, this approach offers a pathway to more robust, interpretable, and accurate language models in biomedical research and clinical practice.
报告人简介：Shankai received his Ph.D. degree in Department of Computer Science at City University of Hong Kong (4th in QS World University Ranking aged under 50 and 5th in Times University Ranking aged under 50) under the supervision of Dr. WONG Ka-Chun. After that, he was awarded the fellowship of NIH and worked as a Postdoctoral Fellow in Zhiyong Lu’s BioNLP Lab at NCBI/NLM/NIH. He is now an associate professor in School of Computer Science and Technology at Hainan University. He has published papers with interesting topics on top conference and journals (e.g. ACL(BioNLP), AMIA, Bioinformatics, FGCS, JBI and JBHI), demonstrating his solid research skills. His research is mainly about applied machine learning in biomedical and clinical fields. His goal is to facilitate knowledge discovery and decision making in healthcare research communities using AI technology.
Shankai has served as the guest editor for the special issue Generative AI in Computational Biology and Bioinformatics for Computational and Structural Biotechnology Journal. He has also served as the co-chair at ISMSI 2024, ISCMI 2024 and the session chair at IEEE UIC 2022, IEEE ICPADS 2023. He has been the program committee member of CACML, BIBM, ICCBB, ICONIP, ACL BioNLP Workshop, etc. He has served as the peer reviewer for multiple conferences (e.g. ACL, BIGCOM, ICONIP, BIBM, ICHI, DMBD, ISMSI, etc.) and journals (e.g. BIB, Bioinformatics, PlosCB, JAMIA, Database, JBI, CSBJ, IEEE TCBB, CAIS, TCSS, ASOC, FGCS, NCAA, etc.).
