Explainable Knowledge-Based Learning for Online Medical Question Answering

This study introduces an explainable AI framework for online medical Question Answering (QA) tasks, a growing need in Internet hospitals and digital healthcare services. In response to the increasing demand for automated yet accountable medical consultation, we develop an expert-curated standard QA pool with vetted medical questions and their corresponding answers. Our work introduces a novel, knowledge-driven learning framework centered around the Attentive Fuzzy Bag-of-Words (AFBoW) model for effective sentence pair matching in medical QA. The proposed two-stage coarse-to-fine sentence matching framework combines a similarity-based search engine and Siamese recurrent neural networks, ensuring a robust and explainable matching process. Experimental results on real-world medical data demonstrate the model's efficacy in enhancing operational efficiency in online medical services while maintaining a high level of explainability.