Exploring Retraining-Free Speech Recognition for Intra-sentential Code-Switching

TL;DR

This paper introduces a retraining-free approach for intra-sentential code-switching speech recognition, utilizing existing models with new pronunciation and language model enhancements, achieving significant error rate reductions.

Contribution

The study presents novel methods for foreign pronunciation generation and language model enrichment that do not require retraining of acoustic or language models.

Findings

Achieved 55.5% relative word error rate reduction

Improved handling of accented foreign pronunciations

Enhanced intra-sentential code-switching recognition performance

Abstract

In this paper, we present our initial efforts for building a code-switching (CS) speech recognition system leveraging existing acoustic models (AMs) and language models (LMs), i.e., no training required, and specifically targeting intra-sentential switching. To achieve such an ambitious goal, new mechanisms for foreign pronunciation generation and language model (LM) enrichment have been devised. Specifically, we have designed an automatic approach to obtain high quality pronunciation of foreign language (FL) words in the native language (NL) phoneme set using existing acoustic phone decoders and an LSTM-based grapheme-to-phoneme (G2P) model. Improved accented pronunciations have thus been obtained by learning foreign pronunciations directly from data. Furthermore, a code-switching LM was deployed by converting the original NL LM into a CS LM using translated word pairs and borrowing statistics for the NL LM. Experimental evidence clearly demonstrates that our approach better deals with accented foreign pronunciations than techniques based on human labeling. Moreover, our best system achieves a 55.5% relative word error rate reduction from 34.4%, obtained with a conventional monolingual ASR system, to 15.3% on an intra-sentential CS task without harming the monolingual recognition accuracy.