基于大语言模型的中文语音识别错误校正中的拼音正则化
Pinyin Regularization in Error Correction for Chinese Speech Recognition with Large Language Models
AI总结：
本文构建中文ASR错误校正基准ChineseHP（724K数据对），提出拼音正则化提示方法，持续提升LLM的中文错误校正能力。
摘要：
近期研究证明了大语言模型（LLM）在自动语音识别（ASR）错误校正中的有效性。然而，大部分研究集中于英语。本文将注意力转向中文。首先，我们构建了一个专门针对中文ASR错误校正的基准数据集，包含724K条假设-转录对，命名为Chinese Hypotheses Paradise数据集（ChineseHP），涵盖广泛场景并具有重大挑战。随后，我们使用该数据集对直接提示和微调预训练LLM进行了初步评估。此外，我们提出了一种直接的拼音正则化提示方法，直接从文本假设转录拼音。实验结果表明，与无正则化相比，拼音正则化持续提升LLM的错误校正能力。数据集已在网站公开。
英文摘要：
Recent studies have demonstrated the efficacy of large language models (LLMs) in error correction for automatic speech recognition (ASR). However, much of the research focuses on the English language. This paper redirects the attention to Chinese. Firstly, we construct a specialized benchmark dataset aimed at error correction for Chinese ASR with 724K hypotheses-transcription pairs, named the Chinese Hypotheses Paradise dataset (ChineseHP), which contains a wide range of scenarios and presents significant challenges. Subsequently, we conduct a preliminary evaluation using the dataset for both direct-prompting and fine-tuning pre-trained LLMs. Furthermore, we propose a straightforward method of Pinyin regularization for prompts, which involves the transcription of Pinyin directly from text hypotheses. The experimental results reveal that Pinyin regularization consistently enhances the error-correcting ability of LLMs when compared with those without regularization. The dataset is available on the website.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
