Retrieval augmented answer generation using M-Files Metadata

Multisource Multi Modal RAG framework: The project will focus on developing a Retrieval-Augmented Generation (RAG) system that can effectively combine and synthesize information from multiple sources within M-Files, utilizing multiple Large Language Models (LLMs) for enhanced accuracy and context. The research will address challenges related to integrating and harmonizing disparate document formats, sources of information and metadata structures, optimizing cross-model communication, evaluation, and ensuring accurate, contextually relevant answer generation while maintaining information integrity and security.

Faculty Supervisor:

Paul Cook

Student:

Partner:

M-Files Canada Inc.

Discipline:

Computer science

Sector:

Professional, scientific and technical services

University:

University of New Brunswick

Program: