REVIEW 2 cited by
Adaptive Contextual Biasing for Transducer Based Streaming Speech Recognition
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Adaptive Contextual Biasing for Transducer Based Streaming Speech Recognition
read the original abstract
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.
Forward citations
Cited by 2 Pith papers
-
Improving Synthetic Data Training for Contextual Biasing Models with a Keyword-Aware Cost Function
A keyword-aware loss with masked cross-entropy and binary gating terms reduces overfitting in synthetic-data training of TCPGen, improving Whisper WER on NSC Part 2 from 14.16% (AGEM baseline) to 11.81%.
-
Efficient Trie-based Biasing using K-step Prediction for Rare Word Recognition
A future-token prediction branch in Whisper gates trie-based biasing rewards, letting greedy decoding recognize rare words without a beam-search reward revocation step.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.