Keyword-optimized Template Insertion for Clinical Information Extraction via Prompt-based Learning
This addresses the challenge of sparse task-relevant information in clinical notes for healthcare NLP practitioners, but it is incremental as it focuses on optimizing template position within an existing prompt-based learning framework.
The paper tackles the problem of clinical note classification with limited annotated data by developing a keyword-optimized template insertion method (KOTI) to improve prompt-based learning, showing performance gains in zero-shot and few-shot settings on several clinical tasks.
Clinical note classification is a common clinical NLP task. However, annotated data-sets are scarse. Prompt-based learning has recently emerged as an effective method to adapt pre-trained models for text classification using only few training examples. A critical component of prompt design is the definition of the template (i.e. prompt text). The effect of template position, however, has been insufficiently investigated. This seems particularly important in the clinical setting, where task-relevant information is usually sparse in clinical notes. In this study we develop a keyword-optimized template insertion method (KOTI) and show how optimizing position can improve performance on several clinical tasks in a zero-shot and few-shot training setting.