Integrating Knowledge into End-to-End Speech Recognition from External Text-Only Data

TL;DR

This paper introduces LST, a novel method to incorporate external text-only data into end-to-end speech recognition models, enabling the use of full sentence context without increasing inference complexity.

Contribution

The paper proposes a two-stage teacher-student learning approach and a causal cloze language model to effectively integrate external text data and utilize sentence context in AED models.

Findings

Improved recognition accuracy on Chinese datasets AISHELL-1 and AISHELL-2.

Effective use of external text-only data without added inference complexity.

Leveraging full sentence context enhances model performance.

Abstract

Attention-based encoder-decoder (AED) models have achieved promising performance in speech recognition. However, because of the end-to-end training, an AED model is usually trained with speech-text paired data. It is challenging to incorporate external text-only data into AED models. Another issue of the AED model is that it does not use the right context of a text token while predicting the token. To alleviate the above two issues, we propose a unified method called LST (Learn Spelling from Teachers) to integrate knowledge into an AED model from the external text-only data and leverage the whole context in a sentence. The method is divided into two stages. First, in the representation stage, a language model is trained on the text. It can be seen as that the knowledge in the text is compressed into the LM. Then, at the transferring stage, the knowledge is transferred to the AED model via teacher-student learning. To further use the whole context of the text sentence, we propose an LM called causal cloze completer (COR), which estimates the probability of a token, given both the left context and the right context of it. Therefore, with LST training, the AED model can leverage the whole context in the sentence. Different from fusion based methods, which use LM during decoding, the proposed method does not increase any extra complexity at the inference stage. We conduct experiments on two scales of public Chinese datasets AISHELL-1 and AISHELL-2. The experimental results demonstrate the effectiveness of leveraging external text-only data and the whole context in a sentence with our proposed method, compared with baseline hybrid systems and AED model based systems.