# PERL: Pinyin Enhanced Rephrasing Language Model for Chinese ASR N-best Error Correction

arXiv: https://arxiv.org/abs/2412.03230 · HTML: https://mlsanity.com/paper/2412.03230/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 17 of its 25 references that are in the map.
- First posted on arXiv 4 Dec 2024.

## Builds on (most cited first, 17 of 17)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Revisiting Pre-Trained Models for Chinese Natural Language Processing](https://mlsanity.com/paper/2004.13922/) (2020)
- [Spelling Error Correction with Soft-Masked BERT](https://mlsanity.com/paper/2005.07421/) (2020)
- [On the (In)Effectiveness of Large Language Models for Chinese Text Correction](https://mlsanity.com/paper/2307.09007/) (2023)
- [Read, Listen, and See: Leveraging Multimodal Information Helps Chinese Spell Checking](https://mlsanity.com/paper/2105.12306/) (2021)
- [N-best T5: Robust ASR Error Correction using Multiple Input Hypotheses and Constrained Decoding Space](https://mlsanity.com/paper/2303.00456/) (2023)
- [MuCGEC: a Multi-Reference Multi-Source Evaluation Dataset for Chinese Grammatical Error Correction](https://mlsanity.com/paper/2204.10994/) (2022)
- [FastCorrect 2: Fast Error Correction on Multiple Candidates for Automatic Speech Recognition](https://mlsanity.com/paper/2109.14420/) (2021)
- [General and Domain Adaptive Chinese Spelling Check with Error Consistent Pretraining](https://mlsanity.com/paper/2203.10929/) (2022)
- [Improving Chinese Spelling Check by Character Pronunciation Prediction: The Effects of Adaptivity and Granularity](https://mlsanity.com/paper/2210.10996/) (2022)
- [Chinese Spelling Correction as Rephrasing Language Model](https://mlsanity.com/paper/2308.08796/) (2023)
- [Multi-stage Large Language Model Correction for Speech Recognition](https://mlsanity.com/paper/2310.11532/) (2023)
- [C-LLM: Learn to Check Chinese Spelling Errors Character by Character](https://mlsanity.com/paper/2406.16536/) (2024)
- [From Spelling to Grammar: A New Framework for Chinese Grammatical Error Correction](https://mlsanity.com/paper/2211.01625/) (2022)
- [Mitigating Catastrophic Forgetting in Multi-domain Chinese Spelling Correction by Multi-stage Knowledge Transfer Framework](https://mlsanity.com/paper/2402.11422/) (2024)

## Built on by (most cited first, 1 of 1)

- [RAIR: Retrieval-Augmented Iterative Refinement for Chinese Spelling Correction](https://mlsanity.com/paper/2504.18938/) (2025)

## Abstract (arXiv)

Chinese ASR correction is challenging because errors are often phonetic (many characters share similar Pinyin) while the correction model must also obey a length constraint under noisy N-best hypotheses. Existing approaches either exploit Pinyin only at the prompt/feature level without integrating it into model representations or rely on generative decoding that can drift in length. We propose PERL, a constrained rephrasing pipeline for Chinese N-best ASR correction that (i) predicts the target length and enforces it via mask budgeting, and (ii) fuses semantic and phonetic (Pinyin) representations through token-wise gates conditioned on sentence semantics. Experiments on Aishell-1 and our new domain N-best benchmark DoAD show that PERL consistently reduces CER (29.11% on Aishell-1 and up to ~70% on DoAD) while maintaining low latency. We also provide analyzes of length generalization and phonetic--semantic interactions, showing when PERL relies on phonetic cues versus semantic constraints.