Saved in:
Bibliographic Details
Main Authors: Matar, Bassam, Fayed, Mohamed, Khalafallah, Ayman
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2509.23504
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911181911883776
author Matar, Bassam
Fayed, Mohamed
Khalafallah, Ayman
author_facet Matar, Bassam
Fayed, Mohamed
Khalafallah, Ayman
contents This paper describes AraS2P, our speech-to-phonemes system submitted to the Iqra'Eval 2025 Shared Task. We adapted Wav2Vec2-BERT via Two-Stage training strategy. In the first stage, task-adaptive continue pretraining was performed on large-scale Arabic speech-phonemes datasets, which were generated by converting the Arabic text using the MSA Phonetiser. In the second stage, the model was fine-tuned on the official shared task data, with additional augmentation from XTTS-v2-synthesized recitations featuring varied Ayat segments, speaker embeddings, and textual perturbations to simulate possible human errors. The system ranked first on the official leaderboard, demonstrating that phoneme-aware pretraining combined with targeted augmentation yields strong performance in phoneme-level mispronunciation detection.
format Preprint
id arxiv_https___arxiv_org_abs_2509_23504
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AraS2P: Arabic Speech-to-Phonemes System
Matar, Bassam
Fayed, Mohamed
Khalafallah, Ayman
Computation and Language
This paper describes AraS2P, our speech-to-phonemes system submitted to the Iqra'Eval 2025 Shared Task. We adapted Wav2Vec2-BERT via Two-Stage training strategy. In the first stage, task-adaptive continue pretraining was performed on large-scale Arabic speech-phonemes datasets, which were generated by converting the Arabic text using the MSA Phonetiser. In the second stage, the model was fine-tuned on the official shared task data, with additional augmentation from XTTS-v2-synthesized recitations featuring varied Ayat segments, speaker embeddings, and textual perturbations to simulate possible human errors. The system ranked first on the official leaderboard, demonstrating that phoneme-aware pretraining combined with targeted augmentation yields strong performance in phoneme-level mispronunciation detection.
title AraS2P: Arabic Speech-to-Phonemes System
topic Computation and Language
url https://arxiv.org/abs/2509.23504