Automatic Speech Recognition and Topic Identification for Almost-Zero-Resource Languages

TL;DR

This paper presents a rapid adaptation approach for universal phone-based ASR systems tailored for almost-zero-resource languages, significantly improving performance in low-resource scenarios.

Contribution

It introduces a Kaldi-based universal phone modeling approach with quick adaptation recipes, advancing low-resource ASR development.

Findings

Outperforms many competing approaches on NIST LoReHLT 2017 datasets

Demonstrates effectiveness of rapid adaptation in extremely low-resource settings

Validates approach within DARPA LORELEI framework

Abstract

Automatic speech recognition (ASR) systems often need to be developed for extremely low-resource languages to serve end-uses such as audio content categorization and search. While universal phone recognition is natural to consider when no transcribed speech is available to train an ASR system in a language, adapting universal phone models using very small amounts (minutes rather than hours) of transcribed speech also needs to be studied, particularly with state-of-the-art DNN-based acoustic models. The DARPA LORELEI program provides a framework for such very-low-resource ASR studies, and provides an extrinsic metric for evaluating ASR performance in a humanitarian assistance, disaster relief setting. This paper presents our Kaldi-based systems for the program, which employ a universal phone modeling approach to ASR, and describes recipes for very rapid adaptation of this universal ASR system. The results we obtain significantly outperform results obtained by many competing approaches on the NIST LoReHLT 2017 Evaluation datasets.