CatDB: Data-catalog-guided, LLM-based Generation of Data-centric ML Pipelines

Summary: CatDB leverages data-catalog metadata to produce dataset-specific instructions that steer LLMs to synthesize end-to-end data-centric ML pipelines (cleaning, augmentation, feature engineering, tuning). Built-in validation and error handling reduce hallucinations and yield AutoML-comparable accuracy with much faster scaling. (summarized by gpt-5-mini on Feb 09 2026)

Paper ID: 14095
Venue: VLDB
Year: 2025
Pagerank: 5.093636e-05
Overall Rank: 10,882 | 25.34%
DOI: 10.14778/3742728.3742754

Incoming Non-self Citations Over Time

No non-self incoming citations found for this paper in this database.

Authors

1. Saeed Fathollahzadeh (Concordia University)
2. Essam Mansour (Concordia University)
3. Matthias Boehm (Technical University of Berlin)

BibTeX Citation

@article{fathollahzadeh_vldb25,
        title = {{CatDB: Data-catalog-guided, LLM-based Generation of Data-centric ML Pipelines}},
        author = {Fathollahzadeh, Saeed and Mansour, Essam and Boehm, Matthias},
        journal = {PVLDB},
        series = {{VLDB} '25},
        volume = {18},
        number = {8},
        pages = {2639--2652},
        doi = {10.14778/3742728.3742754},
        url = {https://doi.org/10.14778/3742728.3742754},
        year = {2025}
}

Incoming Citations (Sorted by Pagerank)

Showing 0 of 0 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank

Outgoing Citations (Sorted by Pagerank)

Showing 21 of 21 cited papers.

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

Rank	Cited Paper	Year	Venue	Pagerank
155	MAD Skills: New Analysis Practices for Big Data	2009	VLDB	0.00028713176
279	Text-to-SQL Empowered by Large Language Models: A Benchmark Evaluation	2024	VLDB	0.00022468369
420	Can Foundation Models Wrangle Your Data?	2023	VLDB	0.00018789852
514	Goods: Organizing Google's Datasets	2016	SIGMOD	0.00017178673
582	ActiveClean: Interactive Data Cleaning For Statistical Modeling	2016	VLDB	0.00016148948
1,004	Democratizing Data Science through Interactive Curation of ML Pipelines	2019	SIGMOD	0.00012701932
1,350	Automating Large-Scale Data Quality Verification	2018	VLDB	0.00011065626
1,756	SystemDS: A Declarative Machine Learning System for the End-to-End Data Science Lifecycle	2020	CIDR	9.8172465e-05
2,147	Nearest Neighbor Classifiers over Incomplete Information: From Certain Answers to Certain Predictions	2021	VLDB	9.0831495e-05
2,315	SANTOS: Relationship-based Semantic Table Union Search	2023	SIGMOD	8.7605101e-05
3,228	Correlation Sketches for Approximate Join-Correlation Queries	2021	SIGMOD	7.6215176e-05
3,387	Data Profiling – A Tutorial	2017	SIGMOD	7.4534663e-05
3,905	Automated Feature Engineering for Algorithmic Fairness	2021	VLDB	7.029145e-05
4,240	LIMA: Fine-grained Lineage Tracing and Reuse in Machine Learning Systems	2021	SIGMOD	6.809685e-05
5,051	Towards Benchmarking Feature Type Inference for AutoML Platforms	2021	SIGMOD	6.385354e-05
5,901	A Scalable AutoML Approach Based on Graph Neural Networks	2022	VLDB	6.0467725e-05
6,928	How do Categorical Duplicates Affect ML? A New Benchmark and Empirical Analyses	2024	VLDB	5.7370426e-05
7,232	Saga: A Scalable Framework for Optimizing Data Cleaning Pipelines for Machine Learning Applications	2023	SIGMOD	5.6659017e-05
8,197	SMARTFEAT: Efficient Feature Construction through Feature-Level Foundation Model Interactions	2024	CIDR	5.4691464e-05
9,437	GIDCL: A Graph-Enhanced Interpretable Data Cleaning Framework with Large Language Models	2024	SIGMOD	5.2687567e-05
9,744	Searching a Database of Source Codes Using Contextualized Code Search	2020	VLDB	5.227679e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	9,155	Adda: Towards Efficient in-Database Feature Generation via LLM-based Agents	2025	SIGMOD
2	8,906	Unveiling Challenges for LLMs in Enterprise Data Engineering	2026	VLDB
3	9,391	Demonstration of DB-GPT: Next Generation Data Interaction System Empowered by Large Language Models	2024	VLDB
4	10,431	AutoDDG: Automated Dataset Description Generation using Large Language Models	2026	SIGMOD
5	8,903	DataLoom: Simplifying Data Loading with LLMs	2024	VLDB
6	4,886	Chat2Data: An Interactive Data Analysis System with RAG, Vector Databases and LLMs	2024	VLDB
7	8,914	CtxPipe: Context-aware Data Preparation Pipeline Construction for Machine Learning	2024	SIGMOD
8	6,101	LLM for Data Management	2024	VLDB
9	10,614	LLM-AutoDP: Automatic Data Processing via LLM Agents for Model Fine-tuning	2026	VLDB
10	13,303	Demonstrating CatDB: LLM-based Generation of Data-centric ML Pipelines	2025	SIGMOD