BLAH8 Project: Bioregulatory Event Extraction Using Large Language Models: A Case Study of Rice Literature
Extracting biological regulation events has long been a focal point in the field of biomedical nature language processing (BioNLP). Existing methods often face challenges such as cascading errors in text mining pipelines and topic limitations from the corpus selection. Fortunately, with the emergence of large language models (LLMs), their robust semantic understanding and extensive knowledge background offer a potential solution to alleviate these issues. Towards this goal, during the Biomedical Linked Annotation Hackathon 8 (BLAH 8), this project explores the feasibility of using LLMs for the task of extracting biological regulation events. Taking rice literature as a case, the results indicate promising performance of LLMs in this task, while also emphasizing several challenges that need to be addressed in future work.
Keywords: Oryza sativa, bioregulatory event, text mining, large language model, prompt engineering