Sato: Contextual Semantic Type Detection in Tables
Summary: Sato detects column semantic types by using table-context signals with column values. A hybrid model fuses deep learning on table corpora, topic modeling, and structured prediction, delivering F1 0.925 (support-weighted) and 0.735 (macro), surpassing prior work. (summarized by gpt-5-nano on Feb 09 2026)
Incoming Non-self Citations Over Time
Authors
- 1. Dan Zhang (University of Massachusetts Amherst)
- 2. Yoshihiko Suhara (Megagon Labs)
- 3. Jinfeng Li (Megagon Labs)
- 4. Madelon Hulsebos (The HEINEKEN Company)
- 5. Çağatay Demiralp (Megagon Labs)
- 6. Wang-Chiew Tan (Megagon Labs)
BibTeX Citation
@article{zhang_vldb20,
title = {{Sato: Contextual Semantic Type Detection in Tables}},
author = {Zhang, Dan and Suhara, Yoshihiko and Li, Jinfeng and Hulsebos, Madelon and Demiralp, Çağatay and Tan, Wang-Chiew},
journal = {PVLDB},
series = {{VLDB} '20},
volume = {13},
number = {11},
pages = {1835--1848},
doi = {10.14778/3407790.3407793},
url = {https://doi.org/10.14778/3407790.3407793},
year = {2020}
}
Incoming Citations (Sorted by Pagerank)
Showing 25 of 25 citing papers.
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 7 of 7 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 65 | Freebase: A Collaboratively Created Graph Database For Structuring Human Knowledge | 2008 | SIGMOD | 0.00038697603 |
| 91 | WebTables: Exploring the Power of Tables on the Web | 2008 | VLDB | 0.00034838835 |
| 94 | Potter's Wheel: An Interactive Data Cleaning System | 2001 | VLDB | 0.00034616103 |
| 317 | Annotating and Searching Web Tables Using Entities, Types and Relationships | 2010 | VLDB | 0.00021402049 |
| 753 | Recovering Semantics of Tables on the Web | 2011 | VLDB | 0.00014336094 |
| 3,640 | Semantic Integration in Heterogeneous Databases Using Neural Networks | 1994 | VLDB | 7.2327067e-05 |
| 6,436 | Synthesizing Type-Detection Logic for Rich Semantic Data Types using Open-source Code | 2018 | SIGMOD | 5.8793793e-05 |
Previous
Page 1 / 1
Next
Semantically Similar Papers
| # | Overall Rank | Paper | Year | Venue |
|---|---|---|---|---|
| 1 | 1,923 | Annotating Columns with Pre-trained Language Models | 2022 | SIGMOD |
| 2 | 3,073 | DeepJoin: Joinable Table Discovery with Pre-trained Language Models | 2023 | VLDB |
| 3 | 6,436 | Synthesizing Type-Detection Logic for Rich Semantic Data Types using Open-source Code | 2018 | SIGMOD |
| 4 | 753 | Recovering Semantics of Tables on the Web | 2011 | VLDB |
| 5 | 10,399 | Retrieve-and-Verify: A Table Context Selection Framework for Accurate Column Annotations | 2026 | SIGMOD |
| 6 | 4,515 | ArcheType: A Novel Framework for Open-Source Column Type Annotation using Large Language Models | 2024 | VLDB |
| 7 | 9,067 | Making Table Understanding Work in Practice | 2022 | CIDR |
| 8 | 2,315 | SANTOS: Relationship-based Semantic Table Union Search | 2023 | SIGMOD |
| 9 | 10,785 | Auto-Test: Learning Semantic-Domain Constraints for Unsupervised Error Detection in Tables | 2025 | SIGMOD |
| 10 | 8,458 | RECA: Related Tables Enhanced Column Semantic Type Annotation Framework | 2023 | VLDB |