Project Overview
Development of sophisticated neural machine translation systems that understand Kurdish's complex morphology and dialectal variations. Our transformer-based models incorporate morphological awareness and cultural context to provide accurate translations between Kurdish and English, Arabic, Turkish, and Persian.
Technologies & Methods
Applications
- Cross-cultural Communication
- Content Localization
- Academic Translation
Related Publications
Neural Machine Translation for Kurdish-English Language Pairs
Dr. Karim Mohammad , Dr. Zainab Hussein (2023)
We present a transformer-based neural machine translation system specifically designed for Kurdish-English translation, incorporating morphological awareness and handling dialectal variations across …
Related Datasets
Kurdish-English Parallel Translation Corpus
Published: May 8, 2023 | Size: 450 MB
High-quality parallel corpus containing 500,000 sentence pairs for Kurdish-English translation, covering multiple domains and ensuring balanced representation of both Sorani and Kurmanji dialects.
Kurdish Morphological Analysis Dataset
Published: April 20, 2023 | Size: 85 MB
Comprehensive morphological analysis dataset containing 100,000 Kurdish words with detailed morphological breakdowns, POS tags, and inflectional information for both Sorani and Kurmanji dialects.
Project Statistics
Research Team
Funding
European Union Horizon Research Grant