Leveraging Synthetic Audio Data for End-to-End Low-Resource Speech Translation

2024/06/25 by Yasmin Moslem, Moslem, Yasmin · 7 citations
Computer Science · #Audio and Speech Processing (eess.AS) #Computation and Language (cs.CL) #FOS: Computer and information sciences #FOS: Electrical engineering #Natural Language Processing Techniques #Sound (cs.SD) #Speech Recognition and Synthesis #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2406.17363

openalex publication_date 2024/06/25 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This paper describes our system submission to the International Conference on Spoken Language Translation (IWSLT 2024) for Irish-to-English speech translation. We built end-to-end systems based on Whisper, and employed a number of data augmentation techniques, such as speech back-translation and noise augmentation. We investigate the effect of using synthetic audio data and discuss several methods for enriching signal diversity.

Cited by

• Kunnafonidilaw ka Cadeau: an ASR dataset of present-day Bambara
• Cost Analysis of Human-corrected Transcription for Predominately Oral Languages
• Synthetic Voice Data for Automatic Speech Recognition in African Languages
• It's Not a Walk in the Park! Challenges of Idiom Translation in Speech-to-text Systems
• Bemba Speech Translation: Exploring a Low-Resource African Language
• GMU Systems for the IWSLT 2025 Low-Resource Speech Translation Shared Task
• KIT's Low-resource Speech Translation Systems for IWSLT2025: System Enhancement with Synthetic Data and Model Regularization

Related

• Improving End-to-end Speech Translation by Leveraging Auxiliary Speech and Text Data
• On Using SpecAugment for End-to-End Speech Translation
• Fine-tuning Whisper on Low-Resource Languages for Real-World Applications
• Open-Source System for Multilingual Translation and Cloned Speech Synthesis
• Leveraging Self-Supervised Models for Automatic Whispered Speech Recognition