Project Overview
A curated overview of verified machine translation research involving Kurdish. The collection covers Northern Kurdish and Central Kurdish translation, parallel news and evaluation corpora, low-resource neural machine translation, and multilingual systems including NLLB. External publications and resources are attributed to their original creators and are not presented as KaiLab-developed systems.
Technologies & Methods
Applications
- Kurdish Translation
- Multilingual Communication
- Machine Translation Evaluation
- Language Technology
Related Publications
Bianet: A Parallel News Corpus in Turkish, Kurdish and English
Duygu Ataman (2018)
Introduces a parallel news corpus for English, Turkish and Northern Kurdish (Kurmanji) and evaluates bilingual and multilingual neural machine translation using Nematus. The work provides an early …
Towards Machine Translation for the Kurdish Language
Sina Ahmadi , Mariam Masoud (2020)
Investigates Sorani Kurdish-English machine translation using OpenNMT LSTM encoder-decoder models and several tokenization strategies. The experiments use Tanzil, TED and translated KurdNet material …
TICO-19: the Translation Initiative for COvid-19
Antonios Anastasopoulos (2020)
Presents a multilingual medical and crisis-domain translation benchmark containing professionally translated and quality-controlled material. The release includes separate English-Sorani and …
Leveraging Multilingual News Websites for Building a Kurdish Parallel Corpus
Sina Ahmadi , Hossein Hassani , Daban Q. Jaff (2022)
Develops manually validated parallel news data for Sorani, Kurmanji and English using multilingual news websites. The associated corpus contains Sorani-Kurmanji, English-Kurmanji and English-Sorani …
Scaling neural machine translation to 200 languages
NLLB Team (2024)
Presents the No Language Left Behind multilingual machine translation system. The released model family explicitly supports Central Kurdish in Arabic script and Northern Kurdish in Latin script among …
PARME: Parallel Corpora for Low-Resourced Middle Eastern Languages
Sina Ahmadi (2025)
Presents community-created parallel corpora for low-resource Middle Eastern language varieties and machine translation experiments using NLLB-600M. Kurdish-related resources include Southern Kurdish, …
Related Datasets
Bianet Parallel News Corpus
Published: January 1, 2018 | Size: Release-specific; see official distribution
Parallel news corpus containing Northern Kurdish (Kurmanji), English and Turkish material. The published study reports 6,486 English-Kurmanji and 7,390 Turkish-Kurmanji pairs; current distributions …
KurdishMT Reproducibility Bundle
Published: January 1, 2020 | Size: Aggregate unique-pair count not verified
Reproducibility resources for Sorani-English machine translation experiments, including repackaged Tanzil, TED and KurdNet material, processed splits, tokenization models, scripts and translation …
TICO-19 Kurdish Translation Benchmark
Published: January 1, 2020 | Size: 3,071 sentence pairs per Kurdish-English variety
Medical and crisis-domain machine translation benchmark with separate English-Sorani and English-Kurmanji sets. Each Kurdish-English pair contains 3,071 aligned sentences: 971 development and 2,100 …
KurdishBLARK InterdialectCorpus
Published: January 1, 2020 | Size: 12,327 Sorani-Kurmanji; 1,797 English-Kurmanji; 650 English-Sorani pairs
Manually validated parallel news corpus containing 12,327 Sorani-Kurmanji, 1,797 English-Kurmanji and 650 English-Sorani aligned pairs.
FLORES-200 Kurdish Evaluation Benchmark
Published: January 1, 2022 | Size: 3,001 designed sentences per language; 2,009 publicly released dev/devtest sentences per Kurdish variety
Multilingual machine translation evaluation benchmark containing Central Kurdish (ckb_Arab) and Northern Kurdish (kmr_Latn). Public Kurdish portions contain 997 development and 1,012 development-test …
NLLB-200 Translation Model Family
Published: January 1, 2022 | Size: Representative release: approximately 600M parameters
Released multilingual neural machine translation model family supporting Central Kurdish (ckb_Arab) and Northern Kurdish (kmr_Latn). The 600M distilled checkpoint is a representative publicly …
PARME Parallel Corpora
Published: January 1, 2025 | Size: Southern Kurdish: 9,797; Laki: 3,418; Hawrami: 7,794; Zazaki: 4,401 aligned records as documented by the project
Community-created parallel resources for low-resource Middle Eastern language varieties. Kurdish-related resources identified by the publication include Southern Kurdish, Laki, Hawrami and Zazaki …