arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2502.04055cs.LG
收藏

评估合成表格数据生成中列之间的逻辑关系

Evaluating Inter-Column Logical Relationships in Synthetic Tabular Data Generation

Department of Engineering, University of Cambridge(剑桥大学工程系);The Alan Turing Institute, London(伦敦艾伦·图灵研究所)

Yunbo Long, Liming Xu, Alexandra Brintrup

AI总结:

本文提出三种评估指标,用于评估合成表格数据中列间逻辑关系的保持情况,并通过实验证明现有方法在保持逻辑一致性方面存在不足,讨论了改进逻辑关系建模的可能路径。

摘要:

当前对合成表格数据的评估主要集中在联合分布建模的质量上,往往忽略了其在保持真实事件序列和列间一致实体关系方面的有效性。本文提出了三种评估指标,用于评估合成表格数据中列间逻辑关系的保持情况。我们通过在真实工业数据集上评估经典和最新生成方法的性能来验证这些指标。实验结果表明,现有方法往往无法严格保持逻辑一致性(例如地理或组织中的层级关系)和依赖性(例如时间序列或数学关系),这些对于保持真实世界表格数据的细粒度真实性至关重要。基于这些见解,本文还讨论了在建模合成表格数据分布时更好地捕捉逻辑关系的可能路径。代码可在https://github.com/Yunbo-max/TabLogicEval获取。

英文摘要:

Current evaluations of synthetic tabular data mainly focus on how well joint distributions are modeled, often overlooking the assessment of their effectiveness in preserving realistic event sequences and coherent entity relationships across columns.This paper proposes three evaluation metrics designed to assess the preservation of logical relationships among columns in synthetic tabular data. We validate these metrics by assessing the performance of both classical and state-of-the-art generation methods on a real-world industrial dataset.Experimental results reveal that existing methods often fail to rigorously maintain logical consistency (e.g., hierarchical relationships in geography or organization) and dependencies (e.g., temporal sequences or mathematical relationships), which are crucial for preserving the fine-grained realism of real-world tabular data. Building on these insights, this study also discusses possible pathways to better capture logical relationships while modeling the distribution of synthetic tabular data. The code is available at https://github.com/Yunbo-max/TabLogicEval.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑