MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202621094609 A) filed by Symbiosis International Deemed University on August 04, 2026, for Hybrid Hindi Grammar Error Detection And Correction System Using Multi-Tier Nlp Pipeline.
Inventors include Dr. Sagarkumar Badhiye; Dr. Pradnya Borkar; Siddharth Pimpalkar; Madhura Yawalkar; and Shreyali Dongre.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: ABSTRACT HYBRID HINDI GRAMMAR ERROR DETECTION AND CORRECTION SYSTEM USING MULTI-TIER NLP PIPELINE The present invention provides a hybrid Hindi grammar error detection and correction system (100) implementing a multi-tier natural language processing pipeline for Hindi text in Devanagari script. The system (100) comprises a text normalization module (130) for Devanagari Unicode normalization and tokenization, a POS tagging module (140) for morphological feature extraction, a rule engine module (150) applying deterministic Unicode-aware linguistic rules for detecting gender agreement, vibhakti agreement, subject-verb conjugation, postposition and punctuation errors, a spelling correction module (160) using a modified Levenshtein distance algorithm adapted for Devanagari matra similarity and phonetics, a stylistic analysis module (170) for selective large language model invocation, an aggregation module (180) for compiling error results and a frontend rendering module (190) for displaying errors with colour-coded inline highlighting and single-click correction. The three-tier architecture achieves comprehensive coverage across seven Hindi error categories with a precision-first strategy minimizing unnecessary API calls. [
Disclaimer: Curated by HT Syndication.
