MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202621097726 A) filed by Mr. Yogesh Bhaurao Desale; Dr. Rahul Gopichand Saner; Dr. Atul Padmakar Khose; and Mr. Venkatesh Bharti on August 12, 2026, for Context-Aware Multi-Layer Linguistic Annotation Framework For Marathi And Low-Resource Language Processing.

Inventors include Mr. Yogesh Bhaurao Desale; Dr. Rahul Gopichand Saner; Dr. Atul Padmakar Khose; and Mr. Venkatesh Bharti.

The application for the patent was published on September 25, 2026, under issue no. 39/2026.

Abstract: ABSTRACT CONTEXT-AWARE MULTI-LAYER LINGUISTIC ANNOTATION FRAMEWORK FOR MARATHI AND LOW-RESOURCE LANGUAGE PROCESSING The present invention relates to a computer-implemented framework for automatically 10 generating and managing multi-layer linguistic annotations for Marathi text. The framework comprises a text preprocessing module, contextual analysis engine, morphology annotation engine, Part-of-Speech annotation engine, named entity annotation engine, syntactic dependency annotation engine, semantic role annotation engine, dialect-specific expression annotation engine, hierarchical annotation engine, 15 conflict detection and resolution modules, confidence scoring module, and version-controlled annotation repository. The framework simultaneously analyzes multiple linguistic dimensions and associates generated annotations with corresponding textual segments while preserving contextual relationships between annotation layers. Conflicting annotations generated by different annotation engines or annotators are 20 detected and resolved using contextual information, linguistic rules, annotation hierarchy, confidence values, and predefined resolution criteria. Confidence scores are generated to identify reliable and uncertain annotations, while version-controlled histories preserve annotation modifications and validation records. The resulting structured annotated corpus is suitable for training and evaluating Natural Language 25 Processing systems, Large Language Models, speech technologies, machine translation systems, and other artificial intelligence applications, particularly for Marathi and other low-resource languages.

Disclaimer: Curated by HT Syndication.