Generating Human Readable Transcript For Automatic Speech Recognition With Pre-trained Language Model

Abstract

Modern Automatic Speech Recognition (ASR) systems can achieve high performance in terms of recognition accuracy. However, a perfectly accurate transcript still can be challenging to read due to disfluency, filter words, and other errata common in spoken communication. Many downstream tasks and human readers rely on the output of the ASR system; therefore, errors introduced by the speaker and ASR system alike will be propagated to the next task in the pipeline. In this work, we propose an ASR post-processing model that aims to transform the incorrect and noisy ASR output into a readable text for humans and downstream tasks. We leverage the Metadata Extraction (MDE) corpus to construct a task-specific dataset for our study. Since the dataset is small, we propose a novel data augmentation method and use a two-stage training strategy to fine-tune the RoBERTa pre-trained model. On the constructed test set, our model outperforms a production two-step pipeline-based post-processing method by

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyliao2021generating

Related papers

• Robust Neural Machine Translation For Clean And Noisy Speech Transcripts (2019)0.00
• Improving Punctuation Restoration For Speech Transcripts Via External Data (2021)4.52
• Asr-generated Text For Language Model Pre-training Applied To Speech Tasks (2022)3.58
• Audio-attention Discriminative Language Model For ASR Rescoring (2019)9.23
• Performance Improvements Of Probabilistic Transcript-adapted ASR With Recurrent Neural Network And Language-specific Constraints (2016)0.00
• Cross-modal ASR Post-processing System For Error Correction And Utterance Rejection (2022)0.00
• Multi-modal Data Augmentation For End-to-end ASR (2018)11.67
• A Two-stage Transliteration Approach To Improve Performance Of A Multilingual ASR (2024)0.00