SWORD:联合优化多智能体工作流与提示词的用户行为模拟框架
Get the story
2026年10月7日,arXiv Multiagent Systems 频道发布一手报道,提出 SWORD(Simulation-driven Workflow and Prompt Optimization with Role-based Design)框架。该框架用于用户行为模拟,核心目标是联合优化多智能体工作流拓扑与自然语言提示词。报道指出,SWORD 仅依赖标量任务指标进行优化,无需领域初始化或任务特定工程。目前公开信息仅涉及该框架的提出及其设计要点,尚无后续实验数据、应用案例或同行评议进展的报道。
Generated from reports · updated 3 hr ago
Timeline
Follow the coverage from different angles.
- arXiv · Multiagent SystemsSWORD:用户行为模拟的联合工作流与提示词优化框架
SWORD(Simulation-driven Workflow and Prompt Optimization with Role-based Design)框架被提出,用于联合优化多智能体工作流拓扑与自然语言提示词,仅依赖标量任务指标,无需领域初始化或任务特定工程。
Heat trend
Current heat 9·Comparable peak 10(Oct 7)·Comparable change over 24 hours –
The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.