SAP 提出 STS：用智能体模拟生成企业数据
原标题：Synthesis Through Simulation: Generating Coherent Enterprise Data via Scalable Agent-System Interaction
AI 摘要
SAP 研究团队提出 Synthesis Through Simulation（STS）数据合成范式，让 LLM 智能体在模拟企业环境中通过调用带策略约束的 API 生成数据，从而在构造上保证结构有效性。其通用填充器 Generalist Populator（GP）在无需数据库 schema 的情况下，于全部十个环境中实现 100% 约束满足和 0.88 平均边际保真度，而统计合成器因需种子数据在七个环境中不适用，具备 schema 权限的智能体在航空环境中有 82% 轨迹失败。团队已开源完整框架、十个环境及生成数据集。
正文节选
Synthesis Through Simulation: Generating Coherent Enterprise Data via Scalable Agent-System Interaction Abstract Tool-calling agents have become central to enterprise AI, yet training and evaluating them at scale remains severely constrained due to business and legal restrictions on enterprise systems, data, and database schemas. Tabular data synthesis offers a natural alternative, but its effectiveness is fundamentally limited by structural validity and schema availability, while procedure-base