Structured review on rag- and multi-agent frameworks: literature overview

Authors: Md Monsur Ali, Abdullah Al Foysal, Siddarth Venkateswaran, Ronald Böck

Abstract:

Chatbots and conversational agents are currently an emerging domain, showing multiple applications. The fundamental technology, mainly Large Language Models, usually has a generic understanding of domain knowledge, however struggling in complex conversations. For this, recently different approaches were developed, increasing the flexibility of such applications, namely RAG and MultiAgent Frameworks. The current manuscripts provides a structured overview on current frameworks, based on the PRISMA guidelines and key factors, evaluating their capabilities, flexibility, and usability. In Part I, a systematic literature review is presented which is complemented by an application-based experimental assessment (Part II) in the same proceeding edition.