Lune

ACL Findings2023

Automatic Table Union Search with Tabular Representation Learning

Xuming Hu, Shen Wang, Xiao Qin, Chuan Lei, Zhengyuan Shen, Christos Faloutsos, Asterios Katsifodimos, George Karypis, Lijie Wen, Philip S. Yu

Abstract

Given a data lake of tabular data as well as a query table, how can we retrieve all the tables in the data lake that can be unioned with the query table ? Table union search constitutes an essential task in data discovery and preparation as it enables data scientists to navigate massive open data repositories. Existing methods identify uniability based on column representations (word surface forms or token embeddings) and column relation represented by column representation similarity. However, the semantic similarity obtained between column representations is often insufficient to reveal latent relational features to describe the column relation between pair of columns and not robust to the table noise. To address these issues, in this paper, we propose a multi-stage self-supervised table union search framework called AUTOTUS, which represents column relation as a vector-column relational representation and learn column relational representation in a multi-stage manner that can better describe column relation for table unionability prediction. In particular, the large language model powered contextualized column relation encoder is updated by adaptive clustering and pseudo label classification iteratively so that the better column relational representation can be learned. Moreover, to improve the robustness of the model against table noises, we propose table noise generator to add table noise to the training table data. Experiments on real-world datasets and synthetic test set augmented with table noise show that AUTOTUS achieves 5.2% performance gain over the SOTA baseline.

Ask Lune

Know this paper in minutes, not an afternoon.

Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.

7 top-tier papers have cited it since. Know what they found before your reviewers do.

Questions to start from

Lune can draw on

the full paper · the top-tier papers that cite it · related work

Get the answer

Free to start. No credit card required.

Lineage

  • OmniMatch: Joinability Discovery in Data Products cites Dataset Discovery in Data Lakes
  • OmniMatch: Joinability Discovery in Data Products cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
  • OmniMatch: Joinability Discovery in Data Products cites Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation Learning
  • Distinctiveness Maximization in Datasets Assemblage cites Dataset Discovery in Data Lakes
  • Distinctiveness Maximization in Datasets Assemblage cites Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation Learning
  • TableFormer: Robust Transformer Modeling for Table-Text Encoding cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
  • Enriching Relations with Additional Attributes for ER cites Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation Learning
  • TabSketchFM: Sketch-Based Tabular Representation Learning for Data Discovery Over Data Lakes cites Dataset Discovery in Data Lakes
  • TabSketchFM: Sketch-Based Tabular Representation Learning for Data Discovery Over Data Lakes cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
  • TabSketchFM: Sketch-Based Tabular Representation Learning for Data Discovery Over Data Lakes cites Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation Learning
  • Revisiting Task-Oriented Dataset Search in the Era of Large Language Models: Challenges, Benchmark, and Solution cites Dataset Discovery in Data Lakes
  • Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation Learning cites Dataset Discovery in Data Lakes
  • Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation Learning cites TaBERT: Pretraining for Joint Understanding of Textual and Tabular Data
Builds onCited byLeft to right by year; the more cited, the higher.

Cited by top-tier papers7

  1. OmniMatch: Joinability Discovery in Data ProductsChristos Koutras, Jiani Zhang, Xiao Qin, Chuan Lei et al.VLDB 2025 · 3 citations
  2. Distinctiveness Maximization in Datasets AssemblageTingting Wang, Shixun Huang, Zhifeng Bao, J. Shane Culpepper et al.WWW 2025 · 3 citations
  3. TabSketchFM: Sketch-Based Tabular Representation Learning for Data Discovery Over Data LakesAamod Khatiwada, Harsha Kokel, Ibrahim Abdelaziz, Subhajit Chaudhury et al.ICDE 2025 · 3 citations
  4. Enriching Relations with Additional Attributes for ERMengyi Yan, Wenfei Fan, Yaoshu Wang, Min XieVLDB 2024 · 1 citation
  5. Revisiting Task-Oriented Dataset Search in the Era of Large Language Models: Challenges, Benchmark, and SolutionZixin Wei, Yucan Guo, Jinyang Li, Xiaolin Han et al.VLDB 2026 · 1 citation

Builds on19

  1. TaBERT: Pretraining for Joint Understanding of Textual and Tabular DataPengcheng Yin, Graham Neubig, Wen-tau Yih, Sebastian RiedelACL 2020 · 417 citations
  2. HybridQA: A Dataset of Multi-Hop Question Answering over Tabular and Textual DataWenhu Chen, Hanwen Zha, Zhiyu Chen, Wenhan Xiong et al.EMNLP Findings 2020 · 205 citations
  3. TableFormer: Robust Transformer Modeling for Table-Text EncodingJingfeng Yang, Aditya Gupta, Shyam Upadhyay, Luheng He et al.ACL 2022 · 145 citations
  4. Semantics-aware Dataset Discovery from Data Lakes with Contextualized Column-based Representation LearningGrace Fan, Jin Wang, Yuliang Li, Dan Zhang et al.VLDB 2023 · 139 citations
  5. Dataset Discovery in Data LakesAlex Bogatu, Alvaro A. A. Fernandes, Norman W. Paton, Nikolaos KonstantinouICDE 2020 · 118 citations

Related papers

  1. LIFTus: An Adaptive Multi-Aspect Column Representation Learning for Table Union SearchErmu Qiu, Jun Gao, Yaofeng Tu, Jingru YangICDE 2025
  2. SANTOS: Relationship-based Semantic Table Union SearchAamod Khatiwada, Grace Fan, Roee Shraga, Zixuan Chen et al.SIGMOD 2023 · 61 citations
  3. DeepJoin: Joinable Table Discovery with Pre-trained Language ModelsYuyang Dong, Chuan Xiao, Takuma Nozawa, Masafumi Enomoto et al.VLDB 2023 · 53 citations
  4. Human-Centered Exploration of Table UnionabilityNina Klimenkova, Sreeram Marimuthu, Roee ShragaVLDB 2026
  5. Novel Table SearchBesat Kassaie, Renée J. MillerICDE 2026
Dusk over the sea between two cliffs drawn in fine vertical lines