Research
MedSpeak: A Knowledge Graph-Aided ASR Error Correction Framework for Spoken Medical QA
arXiv:2602.00981v2 Announce Type: replace-cross Abstract: Spoken question-answering (SQA) systems relying on automatic speech recognition (ASR) often struggle with accurately recognizing medical termi
arXiv:2602.00981v2 Announce Type: replace-cross Abstract: Spoken question-answering (SQA) systems relying on automatic speech recognition (ASR) often struggle with accurately recognizing medical terminology. To this end, we propose MedSpeak, a novel knowledge graph-aided ASR error correction framework that refines noisy transcripts and improves downstream answer prediction by leveraging both semantic relationships and phonetic information encoded in a medical knowledge graph, together with the reasoning power of LLMs. Comprehensive experimental results on benchmarks demonstrate that MedSpeak significantly improves the accuracy of medical term recognition and overall medical SQA performance, establishing MedSpeak as a state-of-the-art solution for medical SQA. The code is available at https://github.com/RainieLLM/MedSpeak.
Related
- Reducing the Offline-Streaming Gap for Unified ASR Transducer with Consistency Regularization
- Utterance-Level Methods for Identifying Reliable ASR-Output for Child Speech
- Learning Chain Of Thoughts Prompts for Predicting Entities, Relations, and even Literals on Knowledge Graphs
- Building evidence-based knowledge bases from full-text literature for disease-specific biomedical reasoning
- KG-Reasoner: A Reinforced Model for End-to-End Multi-Hop Knowledge Graph Reasoning
Source: arXiv cs.AI | 2026-04-28