Publications by authors named "Donghong Pei"
Relation extraction using large language models: a case study on acupuncture point locations.
J Am Med Inform Assoc
November 2024
Article Synopsis
- The study investigates the effectiveness of large language models (LLMs), particularly fine-tuned GPT-3.5, in accurately extracting acupoint location relations from a specific dataset of acupuncture points provided by the WHO.
- The researchers compared various models, focusing on key relation types and found that fine-tuned GPT-3.5 achieved the highest micro-average F1 score of 0.92 in performance metrics.
- The findings highlight the importance of domain-specific fine-tuning in improving model performance for acupuncture, paving the way for better clinical decision support and educational tools in this field.
Objective: To investigate surgical methods and therapeutic effects of giant cell tumor of tendon sheath in finger.
Methods: From July 2002 to December 2010,70 patients with giant cell tumor of tendon sheath in finger which confirmed by operation and pathology,were retrospectively analyzed. There were 29 males,41 females with an average of 42 years (ranged, 16 to 61), and the course of disease ranged form 4 months to 6 years (mean 11 months).