Problem
Research questions and friction points this paper is trying to address.
Synthetic Tabular Data
Inter-Column Constraints
Domain Constraints
Data Validity
Constraint-Aware Synthetic Tabular Data Generation via Inter-Column Constraint Discovery with LLM Agents
🤖 AI Summary
This study addresses the prevalent issue of semantic domain constraint violations in synthetic tabular data by proposing a unified, tool-based repair framework leveraging LLM agents. The method automatically discovers inter-column constraints through machine-executable hypotheses and full-table validation interfaces, integrating counterexample-guided revision with a generator-agnostic post-processing mechanism to achieve zero-violation correction for equations, inequalities, and logical dependencies. Experimental results demonstrate that this framework significantly enhances the detection and repair of complex constraints. Furthermore, it ensures semantic consistency while effectively preserving univariate distributional characteristics and improving utility for downstream tasks, offering a robust solution for high-fidelity synthetic data generation.
📝 Abstract
Generating structurally valid synthetic tabular data remains difficult: outputs with high statistical fidelity and downstream utility can still violate semantically meaningful domain constraints. We study the discovery and enforcement of three complementary inter-column constraint families---equations, linear inequalities, and logical dependencies. Our unified tool-grounded workflow represents all three as machine-executable hypotheses and applies a common interface for full-table validation, deterministic diagnosis, and counterexample-guided revision. A generator-agnostic postprocessor coordinates family-specific repairs on outputs from unchanged tabular generators. Across curated behavioral audits and end-to-end evaluations, the complete workflow improves held-out violation detection over one-shot direct prompting, while postprocessing yields zero measured violations for every retained, applicable constraint, improves downstream utility on most datasets, and largely preserves univariate marginals.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Constraint-Aware Generation
LLM Agents
Inter-Column Constraint Discovery
Generator-Agnostic Postprocessing
Counterexample-Guided Revision
🔎 Similar Papers
No similar papers found.
