TurboBias 2.0: Streaming Context-Biasing for Production-Efficient ASR Systems
2026-08-21 • Artificial Intelligence
Artificial IntelligenceComputation and LanguageMachine LearningSound
AI summaryⓘ
The authors present TurboBias 2.0, a system that helps speech recognition software better understand user-specific words or phrases without slowing down the process. Their approach allows multiple users to have their own custom lists of important words recognized accurately at the same time, even during live streaming. They improved an existing method with faster and more efficient handling using GPUs and batch processing. Tests show that TurboBias 2.0 boosts accuracy for tricky words while keeping the system fast.
Automatic Speech RecognitionContextualizationTransducer-based ASRPhrase BoostingLatencyStreaming InferenceBatched DecodingGPU AccelerationBeam-search Decoding
Authors
Vladimir Bataev, Lilit Grigoryan, Andrei Andrusenko, Nikolay Karpov, Vitaly Lavrukhin, Boris Ginsburg
Abstract
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.