用多个角色AI模拟真实投行，协同炒股提升收益
TradingAgents: Multi-Agents LLM Financial Trading Framework
- 设计多角色LLM代理，分工负责分析、决策与风控
- 实测显示收益更高，风险更低，夏普比率显著改善
- 适合想研究智能投研或多智能体系统的开发者
基于大语言模型（LLMs）的多智能体系统在自动化问题求解方面已取得显著进展。在金融领域，现有工作多聚焦于单智能体处理特定任务或独立采集数据的多智能体框架，而对复现真实交易公司协作机制的潜力仍挖掘不足。TradingAgents提出一种受交易公司启发的新型股票交易框架，包含由LLM驱动的专用角色代理：基本面分析师、情绪分析师、技术分析师及具有不同风险偏好的交易员。框架中设有牛市和熊市研究员评估市场状况，风险管控团队实时监控敞口，交易员则综合辩论结果与历史数据做出决策。通过模拟动态协作的交易环境，该框架有效提升了交易表现。详细架构设计与大规模实验表明，其在累计收益率、夏普比率和最大回撤方面均优于基线模型，凸显了多智能体LLM框架在金融交易中的应用前景。代码开源：https://github.com/TauricResearch/TradingAgents。
原文摘要 · Abstract (English)
Significant progress has been made in automated problem-solving using societies of agents powered by large language models (LLMs). In finance, efforts have largely focused on single-agent systems handling specific tasks or multi-agent frameworks independently gathering data. However, the multi-agent systems' potential to replicate real-world trading firms' collaborative dynamics remains underexplored. TradingAgents proposes a novel stock trading framework inspired by trading firms, featuring LLM-powered agents in specialized roles such as fundamental analysts, sentiment analysts, technical analysts, and traders with varied risk profiles. The framework includes Bull and Bear researcher agents assessing market conditions, a risk management team monitoring exposure, and traders synthesizing insights from debates and historical data to make informed decisions. By simulating a dynamic, collaborative trading environment, this framework aims to improve trading performance. Detailed architecture and extensive experiments reveal its superiority over baseline models, with notable improvements in cumulative returns, Sharpe ratio, and maximum drawdown, highlighting the potential of multi-agent LLM frameworks in financial trading. TradingAgents is available at https://github.com/TauricResearch/TradingAgents.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。