Cross-language phoneme mapping for phonetic search keyword spotting in continuous speech of under-resourced languages

Abstract

As automatic speech recognition-based applications become increasingly common in a wide variety of market segments, there is a growing need to support more languages. However, for many languages, the language resources needed to train speech recognition engines are either limited or completely non-existent, and the process of acquiring or constructing new language resources is both long and costly. This paper suggests a methodology that enables Phonetic Search Keyword Spotting to be implemented in a large speech database of any given under-resourced language using cross-language phoneme mappings to another language. The phoneme mapping enables a speech recognition engine from a sufficiently resourced and well-trained source language to be used for phoneme recognition in the new target language. The keyword search is then performed over a lattice of target language phonemes. Three cross-language phoneme mapping techniques are examined: knowledge-based, data-driven and phoneme recognition performance-based. The results suggest that Phonetic Search Keyword Spotting based on the cross-language phoneme mapping approach proposed herein can serve as a quick initial solution for validating keyword spotting applications in new, under-resourced languages.

Cross-language phoneme mapping for phonetic search keyword spotting using multiple source languages

“…Our previous works [8,9] reviewed existing technologies that utilize cross-language phoneme mapping to enable the use of statistically representative acoustic models from a wellresourced language in order to perform KWS in an underresourced language. Examples for various implementations of this concept were proposed in several recent studies.…”

Improving the effectiveness of voice activation systems with machine learning methods

“…This idea was extended by Wang et al (2015) for pre-training of hybrid speech recognition systems. Tetariy et al (2015) proposed to create a phone mapping between languages (namely, English and Spanish in their experiments) to transfer a model pre-trained on a big dataset for use on a smaller dataset in another language. Compared to these methods, the methods investigated in Chapter 3 do not require a phone transcription for the corpus used for pre-training and do not require the existence of a phone mapping between languages.…”