DeRAGEC: Denoising Named Entity Candidates with Synthetic Rationale for ASR Error Correction

Solee Im, Wonjun Lee, Jinmyeong An, Yunsu Kim, Jungseul Ok +1 more
6/9/2025
cs.CL

Official code: solee0022/deragec (8 stars, last commit 2025-05-24).

Abstract

We present DeRAGEC, a method for improving Named Entity (NE) correction in Automatic Speech Recognition (ASR) systems. By extending the Retrieval-Augmented Generative Error Correction (RAGEC) framework, DeRAGEC employs synthetic denoising rationales to filter out noisy NE candidates before correction. By leveraging phonetic similarity and augmented definitions, it refines noisy retrieved NEs using in-context learning, requiring no additional training. Experimental results on CommonVoice and STOP datasets show significant improvements in Word Error Rate (WER) and NE hit ratio, outperforming baseline ASR and RAGEC methods. Specifically, we achieved a 28% relative reduction in WER compared to ASR without postprocessing. Our source code is publicly available at: https://github.com/solee0022/deragec

View on arXivView PDF

Code Implementations(1)

solee0022/deragecOfficial100%
80PythonMay 21, 20251 years ago

Similar Papers

Generative Annotation for ASR Named Entity Correction

87% match2025

Retrieval-Augmented Self-Taught Reasoning Model with Adaptive Chain-of-Thought for ASR Named Entity Correction

86% match2026

DANCER: Entity Description Augmented Named Entity Corrector for Automatic Speech Recognition

85% match2024

Improving Named Entity Transcription with Contextual LLM-based Revision

82% match2025

Unifying Global and Near-Context Biasing in a Single Trie Pass

82% match2024

PMF-CEC: Phoneme-augmented Multimodal Fusion for Context-aware ASR Error Correction with Error-specific Selective Decoding

81% match2025

Cite this paper

@article{im2025deragec,
  title  = {DeRAGEC: Denoising Named Entity Candidates with Synthetic Rationale for ASR Error Correction},
  author = {Solee Im and Wonjun Lee and Jinmyeong An and Yunsu Kim and Jungseul Ok and Gary Geunbae Lee},
  year   = {2025},
  eprint = {2506.07510},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2506.07510v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2506.07510.svg)](https://codeofpaper.com/paper/2506.07510)

Discussion