Research on Intelligent Agent Technology and Applications Based on Large Models
Wenfu Liu, Wentai Chang, Chuan Shi, Yajing Wang, Ran Hu, Caikun Zhang, Haibo Ouyang · 2024
With the rapid growth of available data, the continuous improvement of computing power, and the increasing maturity of large model technology, AI Agents based on large models are gradually becoming the core of AI research and applications. AI Agents possess characteristics such as learning ability, decision-making capability, adaptability, and initiative, enabling them to automatically make choices and take actions to achieve specified goals. Moreover, by integrating large-scale pre-trained model technology, AI Agents can continuously improve their performance based on interactions. This paper will review the background, concepts, and characteristics of AI Agents, focusing on the analysis of the core capabilities and implementation paradigms of large model-based intelligent agents. In line with industry demands, it constructs an electromagnetic target analysis and recognition intelligent agent application framework based on multi-modal large models, providing a reference for readers interested in the research and development of large model intelligence agents.