AI-Powered Multilingual Content Localization Engine for Skill Courses
Main Article Content
Abstract
This paper presents an AI-powered multilingual content localization engine designed to bridge the language accessibility gap in India's vocational and skilling ecosystem. The proposed system integrates multiple AI technologies including Automatic Speech Recognition (ASR) using Whisper, Neural Machine Translation (NMT) using MarianMT, domain- specific glossary alignment via Retrieval-Augmented Generation (RAG), and Text-to-Speech (TTS) synthesis using Coqui TTS into a unified pipeline for both text-based course content localization and video speech dubbing. The engine targets all 22 scheduled Indian languages and incorporates a modular microservice architecture with LMS integration through FastAPI REST endpoints. Key components include a vocal isolation layer using Spleeter, structured PDF and text parsing, structured assessment localization, and a feedback-driven quality improvement loop. The framework addresses challenges of domain terminology misalignment, regional voice naturalness, audio-video synchronization, and scalability for national-level deployment. Evaluation is designed around BLEU scores for translation, Word Error Rate for transcription, and Mean Opinion Score for TTS naturalness. This work contributes a feasible, cost-efficient, and scalable architecture for automated multilingual localization of vocational training material with significant implications for learner inclusion and certification success across India's diverse linguistic communities
Article Details
Section
COPYRIGHT
Submission of a manuscript implies: that the work described has not been published before, that it is not under consideration for publication elsewhere; that if and when the manuscript is accepted for publication, the authors agree to automatic transfer of the copyright to the publisher.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work
- The journal allows the author(s) to retain publishing rights without restrictions.
- The journal allows the author(s) to hold the copyright without restrictions.
References
[1] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, L. Kaiser, and I. Polosukhin, "Attention is all you need," in Advances in Neural Information Processing Systems, vol. 30, 2017.
[2] J. Tiedemann and S. Thottingal, "OPUS-MT - Building open translation services for the World," in Proc. 22nd Annual Conference of the European Association for Machine Translation, 2020, pp. 479-480
.
[3] A. Kunchukuttan, P. Mehta, and P. Bhattacharyya, "The IIT Bombay English-Hindi Parallel Corpus," in Proc. 11th International Conference on Language Resources and Evaluation, 2018, pp. 2828-2835.
[4] Tiwari, K., Patel, M. (2020). Facial Expression Recognition Using Random Forest Classifier. In: Mathur, G., Sharma, H., Bundele, M., Dey, N., Paprzycki, M. (eds) International Conference on Artificial Intelligence: Advances and Applications 2019. Algorithms for Intelligent Systems. Springer, Singapore. https://doi.org/10.1007/978-981-15-1059-5_15.
[5] A. Radford, J. W. Kim, T. Xu, G. Brockman, C. McLeavey, and I. Sutskever, "Robust speech recognition via large-scale weak supervision," in Proc. 40th International Conference on Machine Learning, 2023.
[6] V. Peddinti, G. Chen, V. Manohar, T. Ko, D. Povey, and S. Khudanpur, "JHU ASpIRE system: Robust LVCSR with time delay neural networks and IVECTOR adaptation," in Proc. IEEE ASRU Workshop, 2015, pp. 539-546.
[7] Shekhawat, V.S., Tiwari, M., Patel, M. (2021). A Secured Steganography Algorithm for Hiding an Image and Data in an Image Using LSB Technique. In: Singh, V., Asari, V.K., Kumar, S., Patel, R.B. (eds) Computational Methods and Data Engineering. Advances in Intelligent Systems and Computing, vol 1257. Springer, Singapore. https://doi.org/10.1007/978-981-15-7907-3_35
[8] Patel, Mayank, Neelam Badi, and Amit Sinhal. "The role of fuzzy logic in improving accuracy of phishing detection system." International Journal of Innovative Technology and Exploring Engineering (2019): 3162-3164.
[9] E. Casanova, J. Weber, C. D. Shulby, A. C. Junior, E. Golge, and M. A. Ponti, "YourTTS: Towards zero-shot multi-speaker TTS and zero- shot voice conversion for everyone," in Proc. 39th International Conference on Machine Learning, 2022.
[10] A. Baby, A. P. Thomas, N. Nishanthi, and H. A. Murthy, "Resources for Indian languages," in Proc. 1st Workshop on Technologies for MT of Low Resource Languages, 2016, pp. 46-54.
[11] P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, H. Kuttler, M. Lewis, W. Yih, T. Rocktaschel, S. Riedel, and D. Kiela, "Retrieval-augmented generation for knowledge-intensive NLP tasks," in Advances in Neural Information Processing Systems, vol. 33, 2020,
pp. 9459-9474.