DataSpace：异构工作区可验证分析的数据代理基准
原标题：DataSpace: Benchmarking Data Agents for Verifiable Analytics over Heterogeneous Workspaces
AI 摘要
香港科技大学团队推出 DataSpace 基准，用于评估数据代理在异构工作区（含数据库、文件、视频等）上的可验证分析能力。该基准包含 410 个跨语言任务和 7439 个工件，并作为 KDD Cup 2026 官方评测基准。评测显示最佳准确率仅 66.34%，且多模态证据整合和跨源连接是主要瓶颈，表明该领域仍有很大提升空间。
正文节选
DataSpace: Benchmarking Data Agents for Verifiable Analytics over Heterogeneous Workspaces Abstract Data agents enable natural-language analytics over organizational workspaces, where relevant evidence may be scattered across databases, structured files, long documents, and multimedia. Existing benchmarks largely isolate structured querying, retrieval, or open-ended analysis, leaving heterogeneous evidence discovery, complete tabular outputs, and deterministic evaluation insufficiently unified.