MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641072264 A) filed by Vellore Institute Of Technology on June 10, 2026, for Ai-Powered Educational Content Translation System With Semantic Validation.
Inventors include M. S. Bhuvanesw; N. Bala Ganesh; and S. Sanjana.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: ABSTRACT AI-POWERED EDUCATIONAL CONTENT TRANSLATION SYSTEM WITH SEMANTIC VALIDATION A computer-implemented method (100) for translating educational content for regional students comprises receiving, via a user interface, an uploaded document containing educational content in a source language (104), extracting printed text from the uploaded document using an optical character recognition engine (206), segmenting the extracted text into meaningful units (212), cleaning and normalizing the segmented text to remove noise (214, 216), translating the cleaned text from the source language to a target regional language using a neural machine translation model (306), applying contextual adjustments to the translated text (310), performing a consistency check on the translated text (312), back-translating the translated text to the source language (314), computing a semantic similarity score between the original extracted text and the back-translated text using a sentence embedding model (316), and outputting the translated text when the semantic similarity score satisfies an acceptance threshold (322).
Disclaimer: Curated by HT Syndication.
