Work

Bilingual ASR model with language identification for Brazilian Portuguese and South-American Spanish

Year: 2023

Type: conference-paper

Abstract: Creating accurate and reliable low-resource automatic speech recognition (ASR) models remains challenging due to limited curated data.This work proposes a bilingual ASR model for Brazilian Portuguese and Latin-American Spanish implemented with the Wav2Vec2.0architecture and trained on multiple speech datasets.It combines Language Identification and Speech Recognition, employing a joint feature encoder and task-specific context encoders.Evaluation in the Multilingual Librispeech dataset demonstrates promising results, with an average accuracy of 75.98% for language identification and a competitive Word Error Rate of 30.45% in a bilingual setting, comparable to the Whisper model.

Source: Anais de XXXV Simpósio Brasileiro de Telecomunicações e Processamento de Sinais

DOI: 10.14209/sbrt.2023.1570916069

Authors: Felipe Farias, William Alberto Cruz Castañeda, Wilmer Lobato, Marcellus Amadeus

Language: English

Topic: Speech Recognition and Synthesis

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: bronze

OpenAlex ID: W4387942868

FWCI: 0.111

Cites: 29

Cited by: 1

Related to: 10