Current large language models (LLMs) have proven useful for analyzing financial data, but most existing models, such as BloombergGPT and FinGPT, lack customization for specific user needs. In this paper, we address this gap by developing FinGPT Search Agents tailored for two types of users: individuals and institutions. For individuals, we leverage Retrieval-Augmented Generation (RAG) to integrate local documents and user-specified data sources. For institutions, we employ dynamic vector databases and fine-tune models on proprietary data. There are several key issues to address, including data privacy, the time-sensitive nature of financial information, and the need for fast responses. Experiments show that FinGPT agents outperform existing models in accuracy, relevance, and response time, making them practical for real-world applications.

翻译：当前的大型语言模型（LLMs）已被证明在分析金融数据方面具有实用价值，但大多数现有模型（如BloombergGPT和FinGPT）缺乏针对特定用户需求的定制化能力。本文通过开发面向两类用户（个人与机构）的定制化FinGPT搜索代理来弥补这一不足。针对个人用户，我们利用检索增强生成（RAG）技术整合本地文档与用户指定的数据源；针对机构用户，我们采用动态向量数据库并对专有数据进行模型微调。研究需解决若干关键问题，包括数据隐私性、金融信息的时效性要求以及快速响应需求。实验表明，FinGPT代理在准确性、相关性与响应时间方面均优于现有模型，具备实际应用价值。