Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Domain-Specific Machine Translation to Translate Medicine Brochures in English to Sorani Kurdish

Mariam Shamal, Hossein Hassani

TL;DR

This study addresses the lack of Kurdish-language medical information by building a domain-specific SMT system to translate English medicine brochures into Sorani Kurdish. Using a 22,940-sentence parallel corpus from 319 brochures and the Moses toolkit, the authors conduct seven data-variant experiments, achieving BLEU scores from 22.65 to 48.93, with post-editing using a medical dictionary further boosting performance (56.87, 31.05, 40.01 for three brochures). Human evaluation by Kurdish-speaking pharmacists, physicians, and users indicates modest but meaningful understandability and accuracy, suggesting practical value for medication safety information. The work provides a valuable resource and baseline for Sorani MT in the medical domain and outlines clear avenues for future improvement, including more data, automation of post-editing, context-aware rules, and neural approaches.

Abstract

Access to Kurdish medicine brochures is limited, depriving Kurdish-speaking communities of critical health information. To address this problem, we developed a specialized Machine Translation (MT) model to translate English medicine brochures into Sorani Kurdish using a parallel corpus of 22,940 aligned sentence pairs from 319 brochures, sourced from two pharmaceutical companies in the Kurdistan Region of Iraq (KRI). We trained a Statistical Machine Translation (SMT) model using the Moses toolkit, conducting seven experiments that resulted in BLEU scores ranging from 22.65 to 48.93. We translated three new brochures to improve the evaluation process and encountered unknown words. We addressed unknown words through post-processing with a medical dictionary, resulting in BLEU scores of 56.87, 31.05, and 40.01. Human evaluation by native Kurdish-speaking pharmacists, physicians, and medicine users showed that 50% of professionals found the translations consistent, while 83.3% rated them accurate. Among users, 66.7% considered the translations clear and felt confident using the medications.

Domain-Specific Machine Translation to Translate Medicine Brochures in English to Sorani Kurdish

TL;DR

This study addresses the lack of Kurdish-language medical information by building a domain-specific SMT system to translate English medicine brochures into Sorani Kurdish. Using a 22,940-sentence parallel corpus from 319 brochures and the Moses toolkit, the authors conduct seven data-variant experiments, achieving BLEU scores from 22.65 to 48.93, with post-editing using a medical dictionary further boosting performance (56.87, 31.05, 40.01 for three brochures). Human evaluation by Kurdish-speaking pharmacists, physicians, and users indicates modest but meaningful understandability and accuracy, suggesting practical value for medication safety information. The work provides a valuable resource and baseline for Sorani MT in the medical domain and outlines clear avenues for future improvement, including more data, automation of post-editing, context-aware rules, and neural approaches.

Abstract

Access to Kurdish medicine brochures is limited, depriving Kurdish-speaking communities of critical health information. To address this problem, we developed a specialized Machine Translation (MT) model to translate English medicine brochures into Sorani Kurdish using a parallel corpus of 22,940 aligned sentence pairs from 319 brochures, sourced from two pharmaceutical companies in the Kurdistan Region of Iraq (KRI). We trained a Statistical Machine Translation (SMT) model using the Moses toolkit, conducting seven experiments that resulted in BLEU scores ranging from 22.65 to 48.93. We translated three new brochures to improve the evaluation process and encountered unknown words. We addressed unknown words through post-processing with a medical dictionary, resulting in BLEU scores of 56.87, 31.05, and 40.01. Human evaluation by native Kurdish-speaking pharmacists, physicians, and medicine users showed that 50% of professionals found the translations consistent, while 83.3% rated them accurate. Among users, 66.7% considered the translations clear and felt confident using the medications.
Paper Structure (21 sections, 6 figures, 3 tables)

This paper contains 21 sections, 6 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Challenges in Kurdish Language Text Processing
  3. Historical Evolution of Medical Translation
  4. Related Work
  5. Domain-Specific Machine Translation
  6. Medicine-Related Machine Translation
  7. Translation into Kurdish Language
  8. Methodology
  9. Data Collection
  10. Preprocessing
  11. Data Cleaning
  12. Text Extraction
  13. Sentence Alignment
  14. Corpus Preparation
  15. Experiments and The Trained Model
  16. ...and 6 more sections

Figures (6)

  • Figure 1: Sample Brochure in English
  • Figure 2: Sample Brochure in Sorani Kurdish
  • Figure 3: Sentence Alignment in InterText Editor
  • Figure 4: Moses Training Overview
  • Figure 5: Professionals' Feedback on the Understandability, Accuracy, and Consistency of the Translations
  • ...and 1 more figures