MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202641001515 A) filed by Amrita Vishwa Vidyapeetham on January 06, 2026, for Multi-Agent Retrieval-Augmented Generation System For Multi-Factor Cohort Discovery And Optimal Matching In Heterogeneous Healthcare Data.
Inventors include Achan, Pradeep Padmakshan; Nair, Namrata; Rangan, P. Venkat; and Narayanan, Sankaran.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: The present disclosure provides a multi-agent evidence fusion system (100) and a method (300) for healthcare cohort analysis. The system (100) includes an orchestrator agent (106) operating on a clinical query through an assertion polarity, a case-based retrieval agent (108) operating on a clinical record through a negation- aware representation (108-2) and a temporal alignment (108-4) via dynamic time warping, a literature retrieval agent (110) operating on a candidate publication through a risk-of-bias indicator (110-2) obtained from a structured bias-assessment framework (110-4), a rank-and-fuse agent (112) operating on a candidate item (206) through an adaptive weighting (202), and a generation unit (114) operating on a provenance record (208). Unlike conventional single-corpus retrieval approaches, this multi- agent fusion approach delivers an explainable and auditable cohort analysis suited to clinical decision support and evidence-based decision-making.
Disclaimer: Curated by HT Syndication.
