Adaptive Contextual Biasing for Transducer Based Streaming Speech Recognition
This addresses a key issue for real-world voice assistants by improving recognition accuracy for both personalized and common words, though it is incremental as it builds on existing contextual biasing methods.
The paper tackles the problem of speech recognition for personalized words degrading performance on common words by proposing an adaptive contextual biasing method that dynamically switches bias lists on and off, achieving up to 6.7% and 20.7% relative reductions in WER and CER compared to the baseline.
By incorporating additional contextual information, deep biasing methods have emerged as a promising solution for speech recognition of personalized words. However, for real-world voice assistants, always biasing on such personalized words with high prediction scores can significantly degrade the performance of recognizing common words. To address this issue, we propose an adaptive contextual biasing method based on Context-Aware Transformer Transducer (CATT) that utilizes the biased encoder and predictor embeddings to perform streaming prediction of contextual phrase occurrences. Such prediction is then used to dynamically switch the bias list on and off, enabling the model to adapt to both personalized and common scenarios. Experiments on Librispeech and internal voice assistant datasets show that our approach can achieve up to 6.7% and 20.7% relative reduction in WER and CER compared to the baseline respectively, mitigating up to 96.7% and 84.9% of the relative WER and CER increase for common cases. Furthermore, our approach has a minimal performance impact in personalized scenarios while maintaining a streaming inference pipeline with negligible RTF increase.