Deep biasing (DB) improves the performance of end-to-end automatic speech
recognition (E2E-ASR) for rare words or contextual phrases using a bias list.
However, most existing methods treat bias phrases as sequences of subwords in a
predefined static vocabulary, which can result in ineffective learning of the
dependencies between subwords. More advanced techniques address this problem by
incorporating additional text data, which increases the overall workload. This
paper proposes a dynamic vocabulary where phrase-level bias tokens can be added
during the inference phase. Each bias token represents an entire bias phrase
within a single token, thereby eliminating the need to learn the dependencies
between the subwords within the bias phrases. This method can be applied to
various architectures because it only extends the embedding and output layers
in common E2E-ASR architectures. Experimental results demonstrate that the
proposed method improves the performance of bias phrases on English and
Japanese datasets.

通过使用动态词汇表中的短语级别偏置令牌，提出的方法改善了英语和日语数据集上偏置短语在端到端自动语音识别中的性能。

上下文化动态词汇的自动语音识别

Contextualized Automatic Speech Recognition with Dynamic Vocabulary

Deep biasing for the Transducer can improve the recognition performance of
rare words or contextual entities, which is essential in practical
applications, especially for streaming Automatic Speech Recognition (ASR).
However, deep biasing with large-scale rare words remains challenging, as the
performance drops significantly when more distractors exist and there are words
with similar grapheme sequences in the bias list. In this paper, we combine the
phoneme and textual information of rare words in Transducers to distinguish
words with similar pronunciation or spelling. Moreover, the introduction of
training with text-only data containing more rare words benefits large-scale
deep biasing. The experiments on the LibriSpeech corpus demonstrate that the
proposed method achieves state-of-the-art performance on rare word error rate
for different scales and levels of bias lists.

通过深度偏置识别稀有单词或上下文实体可以提高 Transducer 的识别性能，这在实际应用中尤为重要，特别是对于流式自动语音识别（ASR）。然而，在深度偏置中采用大规模稀有单词仍然具有挑战性，因为当存在更多干扰者并且偏置列表中存在相似字母序列的单词时，性能会显著下降。在本文中，我们将 Transducers 中的稀有单词的音素和文本信息相结合，以区分发音或拼写相似的单词。此外，训练包含更多稀有单词的仅文本数据有助于大规模深度偏置。在 LibriSpeech 语料库上的实验证明，所提出的方法在不同规模和水平的偏置列表的稀有单词错误率上实现了最先进的性能。

通过音素特征和纯文本数据在流式传导语音变调中改进大规模深度偏差

Improving Large-scale Deep Biasing with Phoneme Features and Text-only  Data in Streaming Transducer

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.

提出了一个基于上下文感知的变形器转录方法的自适应上下文偏置方法，通过使用有偏的编码器和预测器嵌入来执行流媒体预测，动态地开关偏差列表，以适应个性化和常见情况，实验证明相对于基线，它可以在常见情况下减少 WER 和 CER 达到 6.7％和 20.7％，在个性化情况下对性能影响极小，同时保持快速的推理管道。