Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{206673,
        author = {Daya Naik and Nivin K S and Ameer Sawad and Arafath Ali and Badarinadh A P and Shivshankar Pawar},
        title = {Gaana Sampradaya: Preserving Folk Wisdom of Kannada Culture in The Digital Age},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {no},
        pages = {128-136},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=206673},
        abstract = {Preserving folk literature is essential for safeguarding cultural identity, especially in the context of low-resource languages such as Kannada. With the increasing use of artificial intelligence for language translation, it becomes important to evaluate how accurately AI systems represent culturally rich and metaphorical content. Kannada folk songs, in particular, contain deep symbolic meanings, traditional expressions, and emotional context that are difficult for general-purpose translation models to capture accurately. This project focuses on evaluating the semantic accuracy of AI-generated English translations of Kannada folk songs.
Many studies in neural machine translation have focused on improving translation for low-resource languages. However, most of this work is centered on generating translations rather than evaluating how well these translations preserve cultural meaning and semantic accuracy. In particular, there is a lack of reliable methods to compare AI-generated translations with expert interpretations, especially in the case of folk literature. This gap highlights the importance of developing an evaluation-based approach that considers multiple aspects of translation quality. In this work, a structured dataset was created containing original Kannada folk lyrics, AI-generated English translations, and expert-provided interpretations. A similarity-based evaluation model was then designed using TF-IDF cosine similarity, chrF proxy score, and Jaccard word set similarity to examine semantic, structural, and vocabulary-level alignment. The system also includes a manual comparison feature that allows individual translation pairs to be evaluated. The findings indicate that while AI translations are able to convey the general meaning, they often fail to capture cultural nuances and metaphorical expressions. Overall, the proposed framework offers a practical and scalable method for assessing translation quality and contributes to ongoing efforts in cultural preservation and low-resource language research.},
        keywords = {Kannada Folk Songs, Machine Translation, Natural Language Processing, Semantic Evaluation, TF-IDF, Cultural Preservation.},
        month = {July},
        }