Korean Tokenization for Beam Search Rescoring in Speech Recognition

TL;DR

This paper introduces a novel Korean tokenization method with a special SkipTC token to improve language model performance in speech recognition, demonstrating lower error rates on a large-scale dataset.

Contribution

A new Korean tokenization approach with SkipTC token that enhances LM learning and improves ASR accuracy, especially on large-scale datasets.

Findings

Lower word error rate with proposed tokenization

Effective pattern regularization for language models

First ASR results on 7,600h Korean dataset

Abstract

The performance of automatic speech recognition (ASR) models can be greatly improved by proper beam-search decoding with external language model (LM). There has been an increasing interest in Korean speech recognition, but not many studies have been focused on the decoding procedure. In this paper, we propose a Korean tokenization method for neural network-based LM used for Korean ASR. Although the common approach is to use the same tokenization method for external LM as the ASR model, we show that it may not be the best choice for Korean. We propose a new tokenization method that inserts a special token, SkipTC, when there is no trailing consonant in a Korean syllable. By utilizing the proposed SkipTC token, the input sequence for LM becomes very regularly patterned so that the LM can better learn the linguistic characteristics. Our experiments show that the proposed approach achieves a lower word error rate compared to the same LM model without SkipTC. In addition, we are the first to report the ASR performance for the recently introduced large-scale 7,600h Korean speech dataset.