MarketSenseAI 2.0: Enhancing Stock Analysis through LLM Agents

题目： MarketSenseAI 2.0：通过 LLM 代理增强股票分析。
作者： MarketSenseAI 团队等（2025）。
摘要要点： 该工作提出一种代理式系统，让大型语言模型作为模块化分析员，负责从新闻中提取信号、生成交易假设，并与数值预测模型协同。论文展示了 LLM 与定量流水线的混合如何提升信号发现、跨资产泛化能力与可解释性，并在多套数据上验证提高了事件预警能力和回测鲁棒性。
MarketSenseAI 2.0: Enhancing Stock Analysis through LLM Agents
Authors: (listed in arXiv Feb 2025 listings; MarketSenseAI team) — 2025.
Source / arXiv listing (q-fin.CP, Feb 2025).
English summary: MarketSenseAI 2.0 introduces an agent-based system where large language models (LLMs) act as modular analysts: extracting signals from news, generating hypotheses, and coordinating with numeric forecasting models. The paper demonstrates how hybrid LLM + quantitative pipelines improve signal discovery, cross-asset generalization, and explainability for traders and quant teams. Experimental results show enhanced early-warning event detection and improved backtest robustness across multiple datasets.