HealthPrompt: A Zero-shot Learning Paradigm for Clinical Natural Language Processing

Developing clinical natural language systems based on machine learning and deep learning is dependent on the availability of large-scale annotated clinical text datasets, most of which are time-consuming to create and not publicly available. The lack of such annotated datasets is the biggest bottleneck for the development of clinical NLP systems. Zero-Shot Learning (ZSL) refers to the use of deep learning models to classify instances from new classes of which no training data have been seen before. Prompt-based learning is an emerging ZSL technique in NLP where we define task-based templates for different tasks. In this study, we developed a novel prompt-based clinical NLP framework called HealthPrompt and applied the paradigm of prompt-based learning on clinical texts. In this technique, rather than fine-tuning a Pre-trained Language Model (PLM), the task definitions are tuned by defining a prompt template. We performed an in-depth analysis of HealthPrompt on six different PLMs in a no-training-data setting. Our experiments show that HealthPrompt could effectively capture the context of clinical texts and perform well for clinical NLP tasks without any training data.