用多轮对话分段纠错，提升长文本语音识别的准确性和流畅性。
Chain of Correction for Full-text Speech Recognition with Large Language Models
- 分段多轮对话纠错，结合上下文理解提升语义准确性。
- 在ChFT数据集上显著优于基线系统，错误修正率明显提高。
- 适合需要高精度长文本语音转写的场景，如法律、医疗记录。
针对自动语音识别（ASR）输出的全文错误修正问题，大型语言模型（LLM）因其能处理多种错误类型（如标点恢复、逆文本规范化）而备受关注。然而，现有方法在稳定性、可控性、完整性与流畅性方面仍存挑战。本文提出链式纠错（Chain of Correction, CoC），采用多轮对话形式，逐段基于预识别文本和全文上下文进行纠错，以增强语义理解。利用开源的ChFT数据集对预训练LLM进行微调后评估，实验表明，CoC在纠正完整文本ASR输出方面显著优于基线及基准系统。进一步分析了纠错阈值以平衡欠修正与过度重写，并将CoC扩展至超长文本场景，探索了其他信息类型在纠错中的引导作用。
原文摘要 · Abstract (English)
Full-text error correction with Large Language Models (LLMs) for Automatic Speech Recognition (ASR) is attracting increased attention for its ability to address a wide range of error types, such as punctuation restoration and inverse text normalization, across long context. However, challenges remain regarding stability, controllability, completeness, and fluency. To mitigate these issues, this paper proposes the Chain of Correction (CoC), which uses a multi-turn chat format to correct errors segment by segment, guided by pre-recognized text and full-text context for better semantic understanding. Utilizing the open-sourced ChFT dataset, we fine-tune a pre-trained LLM to evaluate CoC's performance. Experiments show that CoC significantly outperforms baseline and benchmark systems in correcting full-text ASR outputs. We also analyze correction thresholds to balance under-correction and over-rephrasing, extrapolate CoC on extra-long ASR outputs, and explore using other types of information to guide error correction.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。