Advanced Neural Machine Translation for Low-Resource
Languages
AI Research Project
Source Text
0 characters
Translation
Translating...
⚠️ Disclaimer: This tool is an AI Research Assistant. Translations should be
verified for critical academic use.
📖 User Guide
1️⃣
Select Languages
Choose your source and target languages from the dropdown menus. We
support Bangla, English, and Marma in different scripts.
2️⃣
Enter Text
Type or paste your text in the source text area. The system accepts
text of various lengths for translation.
3️⃣
Translate
Click the "Translate" button and wait for the AI model to process your
text. Results appear in seconds.
4️⃣
Review Results
Check the translation in the target text area. The system uses
state-of-the-art NMT models for accurate results.
5️⃣
Copy & Use
Use the "Copy Translation" button to copy the result to your clipboard
for use in your work.
6️⃣
Provide Feedback
Help us improve! Click the "Feedback" button to share your experience
and suggestions.
💡 Pro Tips
✓For best results, use complete sentences rather than single words.
✓The system works best with formal and academic language.
✓Use the swap button (⇄) to quickly reverse translation direction.
✓Always verify critical translations with native speakers.
👥 Meet the Research Team
Dedicated researchers working on low-resource language translation
Project
Supervisor
MK
Md Kaviul Hossain
Lecturer, Department of Computer Science and
Engineering (CSE)
Project Supervisor
Student
Researchers
MM
MD Mahedul Hasan Mahe
Lead Full-Stack Developer
Designed the responsive frontend architecture and implemented the FastAPI backend logic.
Responsible for system integration and deployment strategies.
SS
Sumaiya Afroz Sayka
AI & Model Researcher
Conducted MELD dataset curation, fine-tuned the NMT/LLM models, and optimized Marma
tokenizer performance. Expert in neural machine translation architectures.
MR
Meherunessa Ritu
AI & Model Researcher
Conducted MELD dataset curation, fine-tuned the NMT/LLM models, and optimized Marma
tokenizer performance. Specializes in low-resource language processing.
AA
Amena Akter
Backend Engineer
Assisted in server-side API integration and system stability testing. Ensures robust
performance and reliability of the translation service.
If you use this translation system in your research or academic work, please cite it as follows:
BibTeX Format:
@misc{mahe2026marma,
title = {Marma Research Assistant Translator:
A Neural Machine Translation System
for Low-Resource Languages},
author = {Mahe, MD Mahedul Hasan and
Sayka, Sumaiya Afroz and
Ritu, Meherunessa and
Akter, Amena},
year = {2026},
month = {February},
note = {Capstone Project},
howpublished = {Web Application},
institution = {Department of Computer Science},
url = {https://github.com/your-repo/marma-translator}
}
APA Format (7th Edition):
Mahe, M. M. H., Sayka, S. A., Ritu, M., & Akter, A. (2026, February).
Marma Research Assistant Translator: A Neural Machine Translation System for Low-Resource
Languages [Capstone Project].
Department of Computer Science.
IEEE Format:
M. M. H. Mahe, S. A. Sayka, M. Ritu, and A. Akter,
"Marma Research Assistant Translator: A Neural Machine Translation System for Low-Resource
Languages,"
Capstone Project, Dept. Comput. Sci., Feb. 2026.
📄 Project Information
Project Type: Capstone Research Project
Date: February 2026
Technology: Neural Machine Translation (NMT)
Focus: Low-Resource Language Processing
Languages: Bangla, English, Marma
Framework: FastAPI + PyTorch
📋 License: This project is available for academic and research purposes.
Please contact the authors for commercial use permissions.