MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072526 A) filed by Saveetha Engineering College on June 11, 2026, for A Multilingual Generative Ai Model For Low Resource Indian Languages.
Inventor includes S. Hemadharsana.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: ABSTRACT India is home to a rich linguistic diversity, yet many regional and indigenous languages remain underrepresented in modem Artificial Intelligence (AI) systems due to limited digital resources, annotated datasets, and computational support. This research proposes a Multilingual Generative AI Model for Low-Resource Indian Languages aimed at improving language understanding and content generation across multiple Indian languages with scarce linguistic resources. The proposed framework leverages transfer learning, multilingual pretraining, and cross-lingual knowledge sharing to enable effective text generation, translation, summarization, and question-answering capabilities in low-resource languages. A transformerbased architecture is trained using a combination of publicly available multilingual corpora, synthetic data augmentation techniques, and language adaptation strategies to overcome data scarcity challenges. The model incorporates linguistic features specific to Indian languages, including diverse scripts, morphology, and code-mixed language patterns. Experimental evaluation is conducted on selected low-resource Indian languages such as Santali, Bodo, Konkani, and Dogri using standard Natural Language Processing (NLP) benchmarks. Results demonstrate significant improvements in text generation quality, semantic coherence, and translation accuracy compared with existing multilingual baseline models. The proposed system contributes toward reducing the digital language divide, promoting linguistic inclusivity, and enabling broader access to AI-powered services for speakers of low-resource Indian languages. This work highlights the potential of generative AI in preserving linguistic heritage while fostering equitable technological development across India's multilingual ecosystem.
Disclaimer: Curated by HT Syndication.
