SiKDD 2025 news

At this years SiKDD conference, Erik Calcina and his co-authors have been awarded the "Best Paper Award" for the paper titled "Semantic Prompting for Large Language Models in Biomedical Named Entity Recognition". The paper focuses on enhancing biomedical named entity recognition by enriching prompts with semantic descriptions of entity labels, evaluated across zero-shot, few-shot, and fine-tuned scenarios using clinical case reports. Results demonstrated that semantic prompts significantly improved model accuracy in low-supervision settings, with F1 score gains of 4-9 points in few-shot scenarios, though benefits diminished once models were fine-tuned with sufficient training data.

The paper is available online: https://aile3.ijs.si/dunja/SiKDD2025/Papers/IS2024_-_SIKDD_2025_paper_3.pdf