MUMBAI, India, Sept. 14 -- Intellectual Property India has published a patent application (202641101105 A) filed by Varsha E; Vindya Vasini D; and Srirubha R on August 21, 2026, for Medintel: A Retrieval-Augmented Generation Framework For Efficient And Reliable Healthcare Conversational Systems.
Inventors include Varsha E; Vindya Vasini D; and Srirubha R.
The application for the patent was published on September 11, 2026, under issue no. 37/2026.
Abstract: Medlntel: A Retrieval-Augmented Generation Framework for Efficient and Reliable Healthcare Conversational Systems Abstract The project aims to address the growing challenge of information overload in the healthcare domain by leveraging advanced AI-driven retrieval and summarization techniques. With the exponential increase in medical publications, clinicians and researchers face difficulties in staying updated with the latest research findings. This project proposes the implementation of a Retrieval-Augmented Generation (RAG) framework integrated with Spring Al, pgvector, and BioBERT/GPT-based models to automate the extraction, retrieval, and summarization of medical research papers. The system operates through a multi-stage pipeline—beginning with data ingestion and embedding generation, followed by vector-based similarity search using pgvector, and culminating in context-aware summarization through the RAG model. This dual-stage approach ensures that retrieved results are both relevant and contextually accurate. By combining semantic search and generative AT, Medlntel offers a unified, low- latency solution for generating evidence-backed insights directly from research documents. For data privacy and reliability, Medlntel is designed for secure offline deployment using a local repository of research datasets such as PubMed or NIH archives. The inclusion of an interactive dashboard built with React enables users to query topics, view summarized findings, and explore supporting citations, metrics in real time. Compared to traditional manual review processes, this system is expected to achieve a 50-70% reduction in literature review time while significantly improving the precision of retrieved evidence. Through the integration of state-of- the-art Al models, scalable vector databases, and healthcare-specific NLP techniques, Medlntel advances toward an intelligent, automated research assistant for clinical and academic use. The project ultimately aims to enhance evidence-based decision-making, accelerate medical research, and contribute to the creation of more reliable, accessible, and trustworthy healthcare information ecosystems.
Disclaimer: Curated by HT Syndication.
