Fine-Tashkeel: Finetuning Byte-Level Models for Accurate Arabic Text Diacritization
Bashar Al-Rfooh, Gheith Abandah, Rami Al-Rfou
Abstract
Most of previous work on learning diacritization of the Arabic language relied on training models from scratch. In this paper, we investigate how to leverage pre-trained language models to learn diacritization. We finetune token-free pre-trained multilingual models (ByT5) to learn to predict and insert missing diacritics in Arabic text, a complex task that requires understanding the sentence semantics and the morphological structure of the tokens. We show that we can achieve state-of-the-art on the diacritization task with minimal amount of training and no feature engineering, reducing WER by 40%. We release our finetuned models for the greater benefit of the researchers in the community.
Related Papers
Advancing Magnetic Materials Discovery -- A structure-based machine learning approach for magnetic ordering and magnetic moment prediction2025-07-02Prompt Mechanisms in Medical Imaging: A Comprehensive Survey2025-06-28Quantum Reinforcement Learning Trading Agent for Sector Rotation in the Taiwan Stock Market2025-06-26Temporal-Aware Graph Attention Network for Cryptocurrency Transaction Fraud Detection2025-06-26Tabular Feature Discovery With Reasoning Type Exploration2025-06-25A Deep Learning Approach to Identify Rock Bolts in Complex 3D Point Clouds of Underground Mines Captured Using Mobile Laser Scanners2025-06-25Relational Deep Learning: Challenges, Foundations and Next-Generation Architectures2025-06-19Enhancing Forecasting Accuracy in Dynamic Environments via PELT-Driven Drift Detection and Model Adaptation2025-06-17