通过垂直拆分表格提升多跳表检索效果，无需查询时调用大模型。
PEARL: Front-Loading Relational Chains for Multi-Hop Table Retrieval
- 将表格按关联列垂直切分，构建可检索的子表单元
- 离线生成多跳查询，3跳任务下召回率提升最高30.05%
- 适合需要高效多表查询且资源受限的应用场景
尽管大语言模型在表格推理方面表现出色，但由于真实数据的碎片化和关联性结构，相关表格的检索仍具挑战。现有方法通常依赖完整表格表示，忽略了由连接关系引发的跨表语义。我们提出PEARL，一种无需训练的框架，采用基于垂直划分的子表编码方式。该方法在离线阶段通过预识别的连接路径生成多跳查询，并将相关列重组为垂直分割的文档单元，从而实现无需查询时调用大模型的高效多表检索。实验表明，PEARL持续优于现有方法，在3跳查询上R@2指标最高提升30.05%。代码已公开于https://github.com/SOOB2NHO/PEARL。
原文摘要 · Abstract (English)
While large language models (LLMs) have shown strong capabilities in tabular reasoning, retrieving relevant tables remains challenging due to the fragmented and relational structure of real-world data. Existing work typically relies on whole table representations that overlook cross-table semantics induced by join relationships. We propose PEARL, a training-free framework that shifts the paradigm toward vertical partitioning-based sub-table encoding. PEARL augments the retrieval corpus offline by generating multi-hop queries over pre-identified join paths and reorganizing relevant columns into vertically partitioned corpus units, enabling effective multi-table retrieval without query-time LLM inference. Experiments show that PEARL consistently outperforms existing methods, with up to +30.05% gains in R@2 on 3-hop queries. The source code is available at https://github.com/SOOB2NHO/PEARL.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。