Contextual Density Ratio for Language Model Biasing of Sequence to Sequence ASR Systems

TL;DR

This paper introduces a contextual density ratio method to improve end-to-end ASR systems' recognition of named entities, achieving significant accuracy gains without harming overall performance.

Contribution

The paper presents a novel density ratio approach for training and adapting E2E ASR models to better recognize context-specific words like named entities.

Findings

Up to 46.5% relative improvement in named entity recognition

Surpasses shallow fusion baseline by 22.1%

Maintains overall recognition accuracy

Abstract

End-2-end (E2E) models have become increasingly popular in some ASR tasks because of their performance and advantages. These E2E models directly approximate the posterior distribution of tokens given the acoustic inputs. Consequently, the E2E systems implicitly define a language model (LM) over the output tokens, which makes the exploitation of independently trained language models less straightforward than in conventional ASR systems. This makes it difficult to dynamically adapt E2E ASR system to contextual profiles for better recognizing special words such as named entities. In this work, we propose a contextual density ratio approach for both training a contextual aware E2E model and adapting the language model to named entities. We apply the aforementioned technique to an E2E ASR system, which transcribes doctor and patient conversations, for better adapting the E2E system to the names in the conversations. Our proposed technique achieves a relative improvement of up to 46.5% on the names over an E2E baseline without degrading the overall recognition accuracy of the whole test set. Moreover, it also surpasses a contextual shallow fusion baseline by 22.1 % relative.