MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641099364 A) filed by Srinivas University Institute Of Engineering And Technology on August 17, 2026, for Ai-Based Knowledge Graph.
Inventors include Ms. Shravya Suvarna; Dr. Shifana Begum; Karthik Sherigar; Krishnakumar Naik; Nagaraj S Karallingannavar; Pramod Mallikarjuna Malagundi; Fathimathul Harshana; H V Pranathi; and Hari Prasad.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: ABSTRACT AI-BASED KNOWLEDGE GRAPH The growing challenge of information overload in digital learning environments highlights an urgent need for solutions that go beyond conventional text-based AI systems. Traditional conversational chatbots present knowledge in linear text format, failing to convey the relational structure between concepts and increasing cognitive load for learners. The proposed AI-Based Knowledge Graph System dynamically transforms natural language queries, source code, and PDF documents into interactive visual knowledge graphs using Large Language Model inference, Abstract Syntax Tree parsing, and Natural Language Processing pipelines. The system supports three input modalities — query-based concept mapping, AST-driven code visualization, and PDF document intelligence all rendered as spatially navigable node-edge graphs through a unified interface. The system further incorporates a proactive knowledge gap detection mechanism that identifies missing prerequisite concepts from the user's session history and surfaces them as distinctively styled nodes within the live graph. Experimental evaluation across 40 undergraduate participants demonstrated a 35% reduction in concept mapping time, a 42% reduction in debugging resolution time, and a significant improvement in task success rates compared to conventional linear-text AI platforms.
Disclaimer: Curated by HT Syndication.
