The Vicomtech Speech Transcription Systems for the Albayzín 2024 Bilingual Basque-Spanish Speech to Text (BBS-S2T) Challenge
Autores: Juan Manuel Martín Doñas
Fecha: 11.11.2024
Abstract
This paper presents the Vicomtech’s submission to the Albayzín 2024 Bilingual Basque-Spanish Speech-to-Text Challenge, which involves evaluating automatic speech transcription systems on recordings of bilingual Basque Parliament plenary sessions. Seven distinct systems were developed and submitted for evaluation: one primary and six contrastive systems. This work also presents the results obtained by each transcription system on the test subset of the Tuning partition and the evaluation data, in addition to the resources needed by each system for both training and decoding processes.
BIB_text
title = {The Vicomtech Speech Transcription Systems for the Albayzín 2024 Bilingual Basque-Spanish Speech to Text (BBS-S2T) Challenge},
pages = {305-309},
keywds = {
albayz´ın evaluations, speech recognition, end-to- end, self-supervised learning, deep learning.
}
abstract = {
This paper presents the Vicomtech’s submission to the Albayzín 2024 Bilingual Basque-Spanish Speech-to-Text Challenge, which involves evaluating automatic speech transcription systems on recordings of bilingual Basque Parliament plenary sessions. Seven distinct systems were developed and submitted for evaluation: one primary and six contrastive systems. This work also presents the results obtained by each transcription system on the test subset of the Tuning partition and the evaluation data, in addition to the resources needed by each system for both training and decoding processes.
}
date = {2024-11-11},
}