Language Agnostic Data-Driven Inverse Text Normalization

TL;DR

This paper introduces a language-agnostic, data-driven inverse text normalization framework that enhances low-resource language processing by leveraging data augmentation and neural machine translation, maintaining high performance across languages.

Contribution

It proposes a novel language-agnostic ITN approach using data augmentation and neural translation, addressing low-resource language challenges.

Findings

Effective for low-resource languages

Preserves performance for high-resource languages

Utilizes data augmentation and neural translation techniques

Abstract

With the emergence of automatic speech recognition (ASR) models, converting the spoken form text (from ASR) to the written form is in urgent need. This inverse text normalization (ITN) problem attracts the attention of researchers from various fields. Recently, several works show that data-driven ITN methods can output high-quality written form text. Due to the scarcity of labeled spoken-written datasets, the studies on non-English data-driven ITN are quite limited. In this work, we propose a language-agnostic data-driven ITN framework to fill this gap. Specifically, we leverage the data augmentation in conjunction with neural machine translated data for low resource languages. Moreover, we design an evaluation method for language agnostic ITN model when only English data is available. Our empirical evaluation shows this language agnostic modeling approach is effective for low resource languages while preserving the performance for high resource languages.