MUMBAI, India, Oct. 26 -- Intellectual Property India has published a patent application (202521011717 A) filed by S P Jain Education Fz LLC on February 12, 2025, for Framework For Reducing Hallucinations And Enhancing Accuracy In Language Models.
Inventor includes Vikram Chandrahas Pandya.
The application for the patent was published on October 23, 2026, under issue no. 43/2026.
Abstract: The Hallucination Reduction and Accuracy Enhancement Framework 140 improves the reliability and accuracy of language model responses through advanced components. The framework 140 integrates Retrieval Interleaved Generation (RIG) 144, employing a continuous verification loop with real-time fact-checking, iterative data querying, and dynamic response refinement. Additionally, the Graph-Based Retrieval-Augmented Generation (RAG) 146 utilizes hierarchical knowledge graphs for entity-relationship mapping and semantic link validation, ensuring contextual relevance and semantic consistency. The Agentic Maker/Checker Framework enables secure content generation and validation via a dual-agent architecture, where maker agents handle content creation and checker agents enforce compliance and verify facts. The framework 140 is designed for regulated industries, optimizing content generation with compliance protocols and secure audit trails. It supports scalability for knowledge-intensive environments by incorporating adaptive learning pathways and personalized sequencing. The key components 142 include RIG 144, Dynamic Content Enrichment, RAG 146, and the Agentic Maker/Checker Framework for real-time validation and refinement.
Disclaimer: Curated by HT Syndication.
