无需训练即可加速识别关键词，支持主流ASR模型
TurboBias: Universal ASR Context-Biasing powered by GPU-accelerated Phrase-Boosting Tree
- 用GPU加速的词增强树实现通用上下文偏置
- 支持20,000个关键词，解码速度不降反升
- 适合需要快速部署关键词识别的开发者
识别特定关键词是上下文感知自动语音识别（ASR）的关键任务。现有上下文偏置方法普遍存在需额外训练、解码速度慢或限制ASR模型类型的问题。本文提出一种通用上下文偏置框架，兼容主流模型类型：CTC、Transducers和注意力编码器-解码器模型。该框架基于GPU加速的词增强树，在浅融合模式下支持贪婪搜索与束搜索解码，即使面对高达20,000个关键词，解码速度也无明显下降。实验结果表明，该方法在准确率和解码速度上均优于现有开源方案。相关代码已作为NeMo工具包的一部分开源。
原文摘要 · Abstract (English)
Recognizing specific key phrases is an essential task for contextualized Automatic Speech Recognition (ASR). However, most existing context-biasing approaches have limitations associated with the necessity of additional model training, significantly slow down the decoding process, or constrain the choice of the ASR system type. This paper proposes a universal ASR context-biasing framework that supports all major types: CTC, Transducers, and Attention Encoder-Decoder models. The framework is based on a GPU-accelerated word boosting tree, which enables it to be used in shallow fusion mode for greedy and beam search decoding without noticeable speed degradation, even with a vast number of key phrases (up to 20K items). The obtained results showed high efficiency of the proposed method, surpassing the considered open-source context-biasing approaches in accuracy and decoding speed. Our context-biasing framework is open-sourced as a part of the NeMo toolkit.
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。