MUMBAI, India, Oct. 5 -- Intellectual Property India has published a patent application (202641115643 A) filed by Ksherasagar Srivathsa on September 27, 2026, for System And Method For Fail-Safe Adaptive Retrieval Of Clinical Information Using Heterogeneous Knowledge And Memory Stores.
Inventors include Ksherasagar Srivathsa; and Kumaraguru Krishnaraj.
The application for the patent was published on October 02, 2026, under issue no. 40/2026.
Abstract: SYSTEM AND METHOD FOR FAIL-SAFE ADAPTIVE RETRIEVAL OF CLINICAL INFORMATION USING HETEROGENEOUS KNOWLEDGE AND MEMORY STORES A computer-implemented system and method for adaptive retrieval of clinical information from heterogeneous knowledge and memory stores is disclosed. A clinical query and conversational session state are received, and the query is characterized to determine a clinical intent and a confidence state. Based on the characterization and session state, a retrieval mode and corresponding retrieval configuration, including a semantic retrieval count and a relational graph traversal depth, are selected. The system selectively accesses a semantic vector store, a structured clinical knowledge graph, a conversational memory store and an episodic memory store, and assembles retrieved information into a query-specific context for generation of a clinical response. When a query is unrecognized or cannot be reliably characterized, retrieval is constrained to a bounded configuration, preventing escalation to broader retrieval. Interaction-derived information is evaluated for consistency with previously stored memories and selectively admitted into episodic memory asynchronously, without delaying the response.
Disclaimer: Curated by HT Syndication.
