arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2506.04565cs.MAcs.CL
收藏

从独立大语言模型到整合智能:复合AI系统综述

From Standalone LLMs to Integrated Intelligence: A Survey of Compound Al Systems

New Jersey Institute of Technology(新泽西理工学院)

Jiayi Chen, Junyi Ye, Guiling Wang

AI总结:

本文综述了复合AI系统,探讨了其整合大语言模型与外部组件的方法,分析了四种基础范式,并指出规模化、互操作性等挑战及未来研究方向。

摘要:

复合AI系统(CAIS)是一种新兴范式,通过整合大语言模型(LLMs)与外部组件,如检索器、代理、工具和协调器,以克服独立模型在需要记忆、推理、实时接地和多模态理解任务中的局限性。这些系统通过将多个专用模块组合成连贯的工作流,实现更强大且上下文感知的行为。尽管在学术界和工业界日益普及,CAIS领域仍碎片化且缺乏统一的分析、分类和评估框架。本文定义了CAIS的概念,提出了基于组件角色和协调策略的多维分类法,并分析了四种基础范式:检索增强生成(RAG)、LLM代理、多模态LLM(MLLM)和协调。我们回顾了代表性系统,比较了设计权衡,并总结了这些范式中的评估方法。最后,我们指出了关键挑战——包括可扩展性、互操作性、基准测试和协调——并概述了未来研究的前景。本文旨在为研究人员和实践者提供全面的基础,以理解和开发下一代系统级人工智能。

英文摘要:

Compound AI Systems (CAIS) are an emerging paradigm that integrates large language models (LLMs) with external components, including retrievers, agents, tools, and orchestrators, to overcome the limitations of standalone models in tasks requiring memory, reasoning, real-time grounding, and multimodal understanding. These systems enable more capable and context-aware behaviors by composing multiple specialized modules into cohesive workflows. Despite growing adoption in both academia and industry, the CAIS landscape remains fragmented and lacks a unified framework for analysis, taxonomy, and evaluation. In this survey, we define the concept of CAIS, propose a multi-dimensional taxonomy based on component roles and orchestration strategies, and analyze four foundational paradigms: Retrieval-Augmented Generation (RAG), LLM Agents, Multimodal LLMs (MLLMs), and Orchestration. We review representative systems, compare design trade-offs, and summarize evaluation methodologies across these paradigms. Finally, we identify key challenges - including scalability, interoperability, benchmarking, and coordination - and outline promising directions for future research. This survey aims to provide researchers and practitioners with a comprehensive foundation for understanding, developing, and advancing the next generation of system-level artificial intelligence.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑