用三阶段框架让大模型纠错更准，减少幻觉错误
Fewer Hallucinations, More Verification: A Three-Stage LLM-Based Framework for ASR Error Correction
- 先预检测错误，再分步推理修正，最后验证推理过程
- 在AISHELL-1等数据集上，字错误率降低最多21%
- 无需额外训练或数据，适合追求高准确率的语音纠错场景
自动语音识别（ASR）错误纠正旨在修正识别错误的同时保留正确文本。尽管传统方法效果有限，大语言模型（LLM）提供了一种无需训练和标注数据的新范式。然而直接使用LLM会引发幻觉问题，导致对正确文本的误改。为此，我们提出可靠的大语言模型纠错框架（RLLM-CF），包含三个阶段：(1) 错误预检测，(2) 思维链子任务迭代修正，(3) 推理过程验证。该方法无需额外信息或模型微调，确保多轮编程下的纠正正确性。在AISHELL-1、AISHELL-2和LibriSpeech上的实验表明，经本框架增强的GPT-4o模型在字错误率（CER）与词错误率（WER）上分别实现21%、11%、9%和11.4%的相对降低。
原文摘要 · Abstract (English)
Automatic Speech Recognition (ASR) error correction aims to correct recognition errors while preserving accurate text. Although traditional approaches demonstrate moderate effectiveness, LLMs offer a paradigm that eliminates the need for training and labeled data. However, directly using LLMs will encounter hallucinations problem, which may lead to the modification of the correct text. To address this problem, we propose the Reliable LLM Correction Framework (RLLM-CF), which consists of three stages: (1) error pre-detection, (2) chain-of-thought sub-tasks iterative correction, and (3) reasoning process verification. The advantage of our method is that it does not require additional information or fine-tuning of the model, and ensures the correctness of the LLM correction under multi-pass programming. Experiments on AISHELL-1, AISHELL-2, and Librispeech show that the GPT-4o model enhanced by our framework achieves 21%, 11%, 9%, and 11.4% relative reductions in CER/WER.
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。