# CaReAQA: A Cardiac and Respiratory Audio Question Answering Model for Open-Ended Diagnostic Reasoning Source: [CaReAQA: A Cardiac and Respiratory Audio Question Answering Model for Open-Ended Diagnostic Reasoning]() Collection index: [llms.txt]() Data updated: 2025-10-26T04:52:00.639Z Authors: Tsai-Ning Wang, Lin-Lin Chen, Neil Zeghidour, Aaqib Saeed Published: 2025-05-02T11:42:46.000Z ## Abstract Medical audio signals, such as heart and lung sounds, play a crucial role in clinical diagnosis. However, analyzing these signals remains challenging: traditional methods rely on handcrafted features or supervised deep learning models that demand extensive labeled datasets, limiting their scalability and applicability. To address these issues, we propose CaReAQA, an audio-language model that integrates a foundation audio model with the reasoning capabilities of large language models, enabling clinically relevant, open-ended diagnostic responses. Alongside CaReAQA, we introduce CaReSound, a benchmark dataset of annotated medical audio recordings enriched with metadata and paired question-answer examples, intended to drive progress in diagnostic reasoning research. Evaluation results show that CaReAQA achieves 86.2% accuracy on open-ended diagnostic reasoning tasks, outperforming baseline models. It also generalizes well to closed-ended classification tasks, achieving an average accuracy of 56.9% on unseen datasets. Our findings show how audio-language integration and reasoning advances medical diagnostics, enabling efficient AI systems for clinical decision support. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [String-level response similarity on RA-QA Multiple-choice, Discriminative tasks](): String-level response similarity; RA-QA Multiple-choice, Discriminative tasks; BERTScore: 0.85. - [String-level response similarity on RA-QA Global, Discriminative tasks](): String-level response similarity; RA-QA Global, Discriminative tasks; BERTScore: 0.89. - [Discriminative tasks on RA-QA](): Discriminative tasks; RA-QA; Accuracy: 67. - [String-level response similarity on RA-QA Single-Verify, Discriminative tasks](): String-level response similarity; RA-QA Single-Verify, Discriminative tasks; BERTScore: 91. - [Regression on RA-QA](): Regression tasks; RA-QA; MAE: 2.61. - [Multimodal Question Answering on CaReSound](): Multimodal Question Answering; CaReSound; Yes/No Accuracy: 93.12. ## Original sources and implementation [Original paper]() [PDF]()