Robust Acoustic and Semantic Contextual Biasing in Neural Transducers for Speech Recognition
This work addresses the challenge of accurately recognizing rare words in speech recognition systems, which is important for applications like personal assistants, but it appears incremental as it builds on existing attention-based contextual biasing methods.
The paper tackles the problem of improving rare-word recognition in end-to-end automatic speech recognition systems by proposing a contextual biasing approach that uses character representations for acoustic similarity and pretrained language models for semantic context. The result shows relative WER improvements of 4.62%-9.26% on Librispeech and 7.91% on an in-house dataset, with up to 36.80% improvement on tail utterances.
Attention-based contextual biasing approaches have shown significant improvements in the recognition of generic and/or personal rare-words in End-to-End Automatic Speech Recognition (E2E ASR) systems like neural transducers. These approaches employ cross-attention to bias the model towards specific contextual entities injected as bias-phrases to the model. Prior approaches typically relied on subword encoders for encoding the bias phrases. However, subword tokenizations are coarse and fail to capture granular pronunciation information which is crucial for biasing based on acoustic similarity. In this work, we propose to use lightweight character representations to encode fine-grained pronunciation features to improve contextual biasing guided by acoustic similarity between the audio and the contextual entities (termed acoustic biasing). We further integrate pretrained neural language model (NLM) based encoders to encode the utterance's semantic context along with contextual entities to perform biasing informed by the utterance's semantic context (termed semantic biasing). Experiments using a Conformer Transducer model on the Librispeech dataset show a 4.62% - 9.26% relative WER improvement on different biasing list sizes over the baseline contextual model when incorporating our proposed acoustic and semantic biasing approach. On a large-scale in-house dataset, we observe 7.91% relative WER improvement compared to our baseline model. On tail utterances, the improvements are even more pronounced with 36.80% and 23.40% relative WER improvements on Librispeech rare words and an in-house testset respectively.