Towards Foundation Database Models

Summary: Argues for foundation database models: pre-trained, dataset- and task-agnostic models that enable low-overhead transfer to unseen datasets and replace expensive one-off per-task training. Proposes an architecture, shows a prototype feasibility study, and outlines a research roadmap. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID: h2c7ac182ab59d831
Venue: CIDR
Year: 2025
Pagerank: 5.5385053e-05
Overall Rank: 7,385 | 50.35%
DOI: -

Incoming Non-self Citations Over Time

Authors

1. Johannes Wehrstein (Google)
2. Carsten Binnig (Google)
3. Fatma Özcan (Google)
4. Shobha Vasudevan (Google)
5. Yu Gan (Google)
6. Yawen Wang (Google)

BibTeX Citation

@inproceedings{wehrstein_cidr25,
        address = {Amsterdam, Netherlands},
        series = {{CIDR} '25},
        title = {{Towards Foundation Database Models}},
        booktitle = {Proceedings of the {Conference} on {Innovative} {Data} {Systems} {Research}},
        author = {Wehrstein, Johannes and Binnig, Carsten and Özcan, Fatma and Vasudevan, Shobha and Gan, Yu and Wang, Yawen},
        year = {2025}
}

Incoming Citations (Sorted by Pagerank)

Showing 4 of 4 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
7,885	Semantic Integrity Constraints: Declarative Guardrails for AI-Augmented Data Processing Systems	2025	VLDB	5.433531e-05
10,359	Leveraging Query Optimizers to Verify the Soundness of LLM-based Query Rewrites for Real-World Workloads, and More!	2026	CIDR	4.9793485e-05
10,918	Evaluating the Practical Effectiveness of LLM-Driven Index Tuning on Microsoft SQL Server	2026	VLDB	4.9793485e-05
10,971	Verified LLM-Based Query Rewriting for Microsoft SQL Server	2026	VLDB	4.9793485e-05

Outgoing Citations (Sorted by Pagerank)

Showing 15 of 15 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
40	The Case for Learned Index Structures	2018	SIGMOD	0.00046284649
85	Learned Cardinalities: Estimating Correlated Joins with Deep Learning	2019	CIDR	0.00035864347
145	Neo: A Learned Query Optimizer	2019	VLDB	0.0002908188
318	DeepDB: Learn from Data, not from Queries!	2020	VLDB	0.00021167555
329	Can Foundation Models Wrangle Your Data?	2023	VLDB	0.00020858443
362	Bao: Making Learned Query Optimization Practical	2021	SIGMOD	0.00019989474
377	TURL: Table Understanding through Representation Learning	2021	VLDB	0.00019570264
461	An End-to-End Learning-based Cost Estimator	2020	VLDB	0.00017829982
868	Learning Multi-dimensional Indexes	2020	SIGMOD	0.00013354403
1,191	Tsunami: A Learned Multi-dimensional Index for Correlated Data and Skewed Workloads	2021	VLDB	0.00011590153
1,839	Why TPC Is Not Enough: An Analysis of the Amazon Redshift Fleet	2024	VLDB	9.5304799e-05
2,478	Learning a Partitioning Advisor for Cloud Databases	2020	SIGMOD	8.4079121e-05
2,885	Zero-Shot Cost Models for Out-of-the-box Learned Cost Prediction	2022	VLDB	7.9094988e-05
5,840	Pre-training Summarization Models of Structured Datasets for Cardinality Estimation	2022	VLDB	5.9737703e-05
6,416	A Unified Transferable Model for ML-Enhanced DBMS	2022	CIDR	5.7920805e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	318	DeepDB: Learn from Data, not from Queries!	2020	VLDB
2	3,060	Towards a Hands-Free Query Optimizer through Deep Learning	2019	CIDR
3	10,154	Structure-Aware Machine Learning over Multi-Relational Databases	2021	SIGMOD
4	10,739	Database Views as Explanations for Relational Deep Learning	2026	VLDB
5	329	Can Foundation Models Wrangle Your Data?	2023	VLDB
6	13,612	From Big Time Series Forecasting to Foundation Models for Structured Data	2026	VLDB
7	4,741	Machine Learning for Databases	2021	VLDB
8	11,825	Data Management Opportunities for Foundation Models	2022	CIDR
9	4,538	One Model to Rule them All: Towards Zero-Shot Learning for Databases	2022	CIDR
10	6,416	A Unified Transferable Model for ML-Enhanced DBMS	2022	CIDR