发表机构
SAP Labs（思爱普实验室）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对企业AI工具调用智能体训练评估受限问题，提出无模式数据合成范式STS，其通用填充器GP在无数据库模式下实现高保真与约束满足，开源了相关框架、环境及数据集。
AI 中文摘要
工具调用智能体已成为企业AI的核心，但由于企业系统、数据和数据库模式的业务及法律限制，大规模训练和评估它们仍受到严重制约。表格数据合成提供了一种自然的替代方案，但其有效性在根本上受限于结构有效性和模式可用性，而基于过程的方法则存在相反的缺陷，通常缺乏分布保真度，且需要针对每个领域进行专门创作。我们提出**通过模拟进行合成（Synthesis Through Simulation，STS）**，这是一种**无模式（schema-free）**的数据合成范式，其中大型语言模型（LLM）智能体通过在模拟企业环境中执行针对策略强制API的操作来生成数据。由于数据是通过定义什么是有效的同一环境生成的，STS从结构上保证了结构有效性，同时将有效性执行与分布建模解耦，允许分别处理这两个问题。STS的领域无关智能体**通用填充器（Generalist Populator，GP）**解决了分布保真度和合成可扩展性的剩余挑战：GP在**不访问数据库（DB）模式**的情况下，在所有十个环境中实现了**0.88的平均边际保真度**和**100%的约束满足度**，而统计合成器因需要必要的种子数据而无法应用于其中七个环境，且具有模式特权的智能体在航空公司环境中因任务组合脆弱，在紧密耦合的工作流上有82%的轨迹失败。我们在该httpsURL开源了完整框架、所有十个环境及生成的数据集。
英文摘要
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-based approaches yield the opposite weakness, typically lacking distributional fidelity without per-domain authoring. We introduce **Synthesis Through Simulation** (STS), a **schema--free** data synthesis paradigm in which an LLM agent generates data by executing operations against policy-enforcing APIs within simulated enterprise environments. Because data is generated through the same environment that defines what is valid, STS guarantees structural validity by construction while decoupling validity enforcement from distribution modeling, allowing each to be addressed independently. The **Generalist Populator** (GP), STS's domain-agnostic agent, addresses the remaining challenges of distributional fidelity and synthesis scalability: GP achieves **0.88** average marginal fidelity and **100\% constraint satisfaction** across all ten environments *without access to DB schemas*, while statistical synthesizers are inapplicable to seven due to necessary seed data requirements, and schema-privileged agents fail 82\% of trajectories on airline environment's tightly coupled workflows due to brittle task composition. We open-source the full framework, all ten environments, and generated datasets at https://github.com/SAP/synthesis-through-simulation.
