🤖 AI Summary
Consumer health queries are often verbose and semantically misaligned with clinical terminology, hindering intent understanding; meanwhile, existing evaluation datasets suffer from high redundancy and data leakage, compromising benchmark fairness. Method: We propose a medical entity-driven contrastive learning framework that identifies salient medical entities in FAQ pairs to pinpoint query focus and constructs high-quality hard negative samples to improve abstractive summarization accuracy. Contribution/Results: We introduce the first entity-guided contrastive learning mechanism for health query summarization; systematically detect and rectify redundancies and leakage in the MQS dataset series, establishing a revised, fair evaluation benchmark. Our method achieves new state-of-the-art ROUGE-1 scores of 52.85, 43.16, 41.31, and 43.52 on MeQSum and three other major benchmarks. The code and cleaned datasets are publicly released.
📝 Abstract
By summarizing longer consumer health questions into shorter and essential ones, medical question-answering systems can more accurately understand consumer intentions and retrieve suitable answers. However, medical question summarization is very challenging due to obvious distinctions in health trouble descriptions from patients and doctors. Although deep learning has been applied to successfully address the medical question summarization (MQS) task, two challenges remain: how to correctly capture question focus to model its semantic intention, and how to obtain reliable datasets to fairly evaluate performance. To address these challenges, this article proposes a novel medical question summarization framework based on entity-driven contrastive learning (ECL). ECL employs medical entities present in frequently asked questions (FAQs) as focuses and devises an effective mechanism to generate hard negative samples. This approach compels models to focus on essential information and consequently generate more accurate question summaries. Furthermore, we have discovered that some MQS datasets, such as the iCliniq dataset with a 33% duplicate rate, have significant data leakage issues. To ensure an impartial evaluation of the related methods, this article carefully examines leaked samples to reorganize more reasonable datasets. Extensive experiments demonstrate that our ECL method outperforms the existing methods and achieves new state-of-the-art performance, i.e., 52.85, 43.16, 41.31, 43.52 in terms of ROUGE-1 metric on MeQSum, CHQ-Summ, iCliniq, HealthCareMagic dataset, respectively. The code and datasets are available at https://github.com/yrbobo/MQS-ECL
