The method for extracting meeting document information in smart meeting rooms uses artificial intelligence to generate enhanced retrieval based on meeting context.

Abstract

This invention proposes a method for extracting meeting document information in smart meeting rooms using artificial intelligence-based meeting-contextualized retrieval augmented generation (MC-RAG). The method utilizes edge devices in conjunction with a facilitator node to capture and convert conversational data into text, perform semantic analysis to create a contextual summary, and accurately identify relevant documents before retrieval. The system employs a distributed vector database, a contextual query rewriting mechanism, and a ranking algorithm to improve accuracy. The solution reduces latency, optimizes computational resources, increases deployment capabilities across edge devices, and enhances the efficiency of information retrieval in smart meetings.