TurboBias 2.0 实现高效个性化语音识别，支持实时流式处理。
TurboBias 2.0: Streaming Context-Biasing for Production-Efficient ASR Systems
- 基于转录模型构建分词级无感提升图，支持多用户独立上下文配置
- 单批内每句独立使用上下文，推理延迟低、吞吐高
- 适用于生产环境，兼容贪婪与束搜索解码，支持离线和流式
上下文增强对生产级自动语音识别（ASR）系统至关重要，需在严格延迟约束下准确识别用户提供的短语。尽管许多上下文偏置方法提升了识别准确率，却常忽视现代生产系统的实际需求：流式推理、高效批量解码、用户专属上下文列表及低运行开销。本文提出 TurboBias 2.0，一个面向生产的基于转换器的 ASR 系统中高效短语增强框架。该框架扩展了 GPU 加速的 TurboBias，引入大小写不敏感的提升图与每流批量解码机制，使批处理中的每条语音可独立使用上下文偏置配置。这实现了多用户并发场景下的个性化上下文增强，且不共享或混淆用户上下文。该框架支持离线与流式推理，兼容贪婪解码与束搜索。实验表明，TurboBias 2.0 在保持低延迟与高吞吐的同时显著提升上下文短语识别准确率。
原文摘要 · Abstract (English)
Contextualization is essential for production automatic speech recognition (ASR) systems, where user-provided phrases must be recognized accurately under strict latency constraints. Although many context-biasing methods improve recognition accuracy, they often do not address the practical requirements of modern production ASR systems: streaming inference, efficient batched decoding, user-specific context lists, and low runtime overhead. We propose TurboBias 2.0, a production-oriented framework for efficient phrase boosting in Transducer-based ASR systems. The framework extends GPU-accelerated TurboBias with a case-insensitive boosting graph and per-stream batched decoding, allowing each utterance in a batch to use an independent context-biasing configuration. This enables personalized context biasing for multiple simultaneous users without sharing or mixing their context lists. The proposed framework supports both offline and streaming inference and can be used with greedy and beam-search decoding. Experiments show that TurboBias 2.0 improves contextual phrase recognition while preserving low latency and high throughput.
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。