A Novel Model For Transliterating MODI To Devanagari Script
DOI:
https://doi.org/10.29284/s06egq94Keywords:
MODI Script Transliteration, Encoder-Decoder Model, LSTM, Deep Sequence Learning, Sequence-to-Sequence Model.Abstract
Preservation and digitization of historical documents inscribed in low-resource scripts continue to pose significant threats in computational linguistics and digital heritage conservation. The MODI script, which has been historically utilized for administrative and literary purposes, is now largely out of the way due to the absence of robust transliteration systems. In the current piece of research work, we present an Encoder-Decoder Deep Sequence (EDDS) model based on stacked Long Short-Term Memory (LSTM) networks for sentence-level transliteration into Devanagari script of MODI script. The proposed model handles variable-length input and output sequences and effectively captures long-range contextual dependencies. A parallel corpus of 37,796 aligned MODI-Devanagari sentence pairs has employed to train and assess the proposed model. Experimental results determine a BLEU score of 46, character-level accuracy of 85.16%, word-level accuracy of 86.12%, and a Character Error Rate (CER) of 0.09, signifying strong transliteration fidelity. The model further achieves an overall transliteration accuracy of 97.16%, demonstrating strong generalization capabilities to unseen and irregular input sequences. These findings validate the effectiveness of deep sequence learning for low-resource historical scripts and highlight the proposed EDDS model as a scalable and reliable solution for MODI-to-Devanagari transliteration.
References
[1] Felix Stahlberg, “Neural Machine Translation: A Review”, Journal of Artificial Intelligence Research, vol. 69, pp. 343-418, 2020.
[2] S. Joseph & J. George, “Machine Transliteration of Handwritten MODI Script to Devanagari using Deep Neural Networks”, International Journal of Computing, Vol. 23, Issue 2, pp. 219-225, June 2024.
[3] Kundaikar, T., Fadte, S., Karmali, R., Wagh, R., Pawar, J.D., “Modi document transcription to Devanagari”, Ingénierie des Systèmes d’Information, Vol. 29, No. 3, pp. 909-915, 2024.
[4] Ilya Sutskever, Oriol Vinyals, and Quoc V. Le., “Sequence to sequence learning with neural networks”, In Proceedings of the 28th International Conference on Neural Information Processing Systems, Vol. 2, pp. 3104-3112, 2014.
[5] K. Kaur and P. Singh, “Review of Machine Transliteration Techniques,” International Journal of Computing Applications, Vol. 107, No. 20, pp. 13-16, 2014.
[6] S. Kundu, S. Paul, and S. Pal, “A Deep Learning Based Approach to Transliteration”, Proceedings of the Seventh Named Entities Workshop, pp. 79-83, 2018.
[7] Bajiyo Baiju, Kavya Manohar, Leena G. Pillai, and Elizabeth Sherly, “Romanized to Native Malayalam Script Transliteration Using an Encoder-Decoder Framework”, In Proceedings of the First Workshop on Natural Language Processing for Indo-Aryan and Dravidian Languages, pp. 174-178, 2024.
[8] Nagaraj, P.K., Ravikumar, K.S., Kasyap, M.S., Murthy, M.H.S., and Paul, J., “Kannada to English machine translation using deep neural network”, Ingénierie des Systèmes d’Information, Vol. 26, No. 1, pp. 123-127, 2021.
[9] Sawant, S., Sharma, A., Suvarna, G., Tanna, T., Kulkarni, S., “Word transcription of MODI script to Devanagari using deep neural network”, In 2020 3rd International Conference on Communication System, Computing and IT Applications, 2020.
[10] Ehud Reiter, “A Structured Review of the Validity of BLEU”, Computational Linguistics, Vol. 44 (3), pp. 393-401, 2018.
[11] Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu., “Bleu: a Method for Automatic Evaluation of Machine Translation”, In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, Philadelphia, Pennsylvania, USA, pp. 311-318, 2002.
[12] Abhinav Jha and Hemprasad Yashwant Patil, “A Review of Machine Transliteration, Translation, Evaluation Metrics and Datasets in Indian Languages”, Multimedia Tools Application 82, 23509-23540, 2023.
[13] Umer Butt, Stalin Varanasi, and Günter Neumann, “Low-Resource Transliteration for Roman-Urdu and Urdu Using Transformer-Based Models”, In Proceedings of the Eighth Workshop on Technologies for Machine Translation of Low-Resource Languages (LoResMT 2025), pp. 144-153, April 2025.
[14] Clemens Neudecker, Konstantin Baierer, Mike Gerber, Christian Clausner, Apostolos Antonacopoulos, and Stefan Pletschacher, “A survey of OCR evaluation tools and metrics”, In Proceedings of the 6th International Workshop on Historical Document Imaging and Processing (HIP '21). Association for Computing Machinery, New York, pp. 13-18, 2021.
[15] PramukhIME software https://www.pramukhime.com/type/marathimodi
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
