Enhancing Code-switching Speech Recognition with Interactive Language Biases

TL;DR

This paper introduces an interactive language biasing approach to improve code-switching speech recognition by leveraging multi-level language information, resulting in better performance on multilingual datasets.

Contribution

It proposes a novel method of biasing hybrid CTC/attention ASR models with multi-level language posteriors and explores their interactions, enhancing code-switching recognition.

Findings

Interactive language biases improve ASR performance on code-switching data.

Different resolutions of language biases interact to influence recognition accuracy.

Language bias implicitly enhances internal language modeling, affecting external language model integration.

Abstract

Languages usually switch within a multilingual speech signal, especially in a bilingual society. This phenomenon is referred to as code-switching (CS), making automatic speech recognition (ASR) challenging under a multilingual scenario. We propose to improve CS-ASR by biasing the hybrid CTC/attention ASR model with multi-level language information comprising frame- and token-level language posteriors. The interaction between various resolutions of language biases is subsequently explored in this work. We conducted experiments on datasets from the ASRU 2019 code-switching challenge. Compared to the baseline, the proposed interactive language biases (ILB) method achieves higher performance and ablation studies highlight the effects of different language biases and their interactions. In addition, the results presented indicate that language bias implicitly enhances internal language modeling, leading to performance degradation after employing an external language model.