Multilingual Text Style Transfer: Datasets & Models for Indian Languages

S Mukherjee, AK Ojha, A Bansal, D Alok… - arXiv preprint arXiv …, 2024 - arxiv.org
Text style transfer (TST) involves altering the linguistic style of a text while preserving its core
content. This paper focuses on sentiment transfer, a vital TST subtask (Mukherjee et al …

[HTML][HTML] Relating pronunciation distance metrics to intelligibility across English accents

T Bent, M Henry, RF Holt, H Lind-Combs - Journal of Phonetics, 2024 - Elsevier
Unfamiliar accents can cause word recognition challenges, particularly in noisy
environments, but few studies have incorporated quantitative pronunciation distance metrics …

Continued Pretraining for Domain Adaptation of Wav2vec2. 0 in Automatic Speech Recognition for Elementary Math Classroom Settings

AA Attia, D Demszky, T Ogunremi, J Liu… - arXiv preprint arXiv …, 2024 - arxiv.org
Creating Automatic Speech Recognition (ASR) systems that are robust and resilient to
classroom conditions is paramount to the development of AI tools to aid teachers and …

Exploring the Impact of Data Quantity on ASR in Extremely Low-resource Languages

YF Cheng, LW Chen, HS Lee, HM Wang - arXiv preprint arXiv:2409.08872, 2024 - arxiv.org
This study investigates the efficacy of data augmentation techniques for low-resource
automatic speech recognition (ASR), focusing on two endangered Austronesian languages …

Improving Access to Untranscribed Speech Corpora Using AI

NM San - 2024 - search.proquest.com
How easily speech corpora can be indexed and searched has a direct impact on how
effectively its contents can be used by many interested parties—from linguists, to language …

Mesure du niveau de proximité entre enregistrements audio et évaluation indirecte du niveau d'abstraction des représentations issues d'un grand modèle de langage

M Fily, G Wisniewski, S Guillaume, G Adda… - Actes de JEP-TALN …, 2024 - inria.hal.science
Nous explorons les représentations vectorielles de la parole à partir d'un modèle pré-
entraîné pour déterminer leur niveau d'abstraction par rapport au signal audio. Nous …

[PDF][PDF] Wave to Interlingua: Analyzing Representations of Multilingual Speech Transformers for Spoken Language Translation

BM Abdullah, MM Shaik, D Klakow - isca-archive.org
Abstract In Transformer-based Speech-to-Text (S2T) translation, an encoder-decoder model
is trained end-to-end to take as input an untranscribed acoustic signal in the source …