Large language models, particularly fine-tuned GPT-3.5, demonstrate superior performance in extracting complex spatial relationships between acupuncture points and human anatomy compared to traditional deep learning models.
Developing a high-throughput biomedical relation extraction system using large language models for semi-structured web articles.
Addressing the challenges of noisy audio in biomedical NER tasks through innovative dataset creation and GPT4-based transcript cleaning.
KnowDDI combines rich biomedical knowledge with deep learning techniques to predict drug-drug interactions accurately and interpretably.
N-ary relation extraction improves performance of Binding events in pipelined biomedical event extraction.
The author presents a novel deep learning framework that combines a Vision Transformer and a bidirectional GRU to enhance glaucoma diagnosis accuracy using 3D OCT imaging.
The author proposes DOSA-MO, a novel multi-objective optimization wrapper algorithm that adjusts performance expectations during optimization to improve model selection and reduce overestimation.