Pseudo-Labeling for Domain-Agnostic Bangla Automatic Speech Recognition

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

One of the major challenges for developing automatic speech recognition (ASR) for low-resource languages is the limited access to labeled data with domain-specific variations. In this study, we propose a pseudo-labeling approach to develop a large-scale domain-agnostic ASR dataset. With the proposed methodology, we developed a 20k+ hours labeled Bangla speech dataset covering diverse topics, speaking styles, dialects, noisy environments, and conversational scenarios. We then exploited the developed corpus to design a conformer-based ASR system. We bench-marked the trained ASR with publicly available datasets and compared it with other available models. To investigate the efficacy, we designed and developed a human-annotated domain-agnostic test set composed of news, telephony, and conversational data among others. Our results demonstrate the efficacy of the model trained on psuedo-label data for the designed test-set along with publicly-available Bangla datasets. The experimental resources will be publicly available.

Original languageEnglish
Title of host publicationBLP 2023 - 1st Workshop on Bangla Language Processing, Proceedings of the Workshop
EditorsFarig Sadeque, Ruhul Amin, Sudipta Kar, Shammur Absar Chowdhury, Firoj Alam
PublisherAssociation for Computational Linguistics (ACL)
Pages196-200
Number of pages5
ISBN (Electronic)9798891760585
Publication statusPublished - 2023
Event1st Workshop on Bangla Language Processing, BLP 2023 - Singapore, Singapore
Duration: 7 Dec 2023 → …

Publication series

NameBLP 2023 - 1st Workshop on Bangla Language Processing, Proceedings of the Workshop

Conference

Conference1st Workshop on Bangla Language Processing, BLP 2023
Country/TerritorySingapore
CitySingapore
Period7/12/23 → …

Fingerprint

Dive into the research topics of 'Pseudo-Labeling for Domain-Agnostic Bangla Automatic Speech Recognition'. Together they form a unique fingerprint.

Cite this