TransClean: Finding False Positives in Multi-Source Entity Matching under Real-World Conditions via Transitive Consistency

2025-06-04

AI Summary

In multi-source entity matching, large-scale, noisy, unlabeled data with distribution drift leads to rampant false positives. Method: This paper proposes an unsupervised false-positive detection framework grounded in transitive consistency—modeling it as a surrogate metric for matching quality. The approach constructs a heterogeneous graph over multi-source records, performs implicit pairwise reasoning, and applies transitive closure analysis to jointly validate matches and localize suspicious record groups—all without human annotations. It integrates mainstream matching models (e.g., DistilBERT, CLER) and supports iterative refinement. Contribution/Results: Evaluated on multi-source entity matching tasks, the framework achieves an average F1-score improvement of 24.42%, demonstrates model-agnosticism and strong robustness to noise and distribution shift, and significantly enhances both true-positive retention and false-positive detection rates.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Machine Learning: Multi-instance/Multi-view Learning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

We present TransClean, a method for detecting false positive predictions of entity matching algorithms under real-world conditions characterized by large-scale, noisy, and unlabeled multi-source datasets that undergo distributional shifts. TransClean is explicitly designed to operate with multiple data sources in an efficient, robust and fast manner while accounting for edge cases and requiring limited manual labeling. TransClean leverages the Transitive Consistency of a matching, a measure of the consistency of a pairwise matching model f_theta on the matching it produces G_f_theta, based both on its predictions on directly evaluated record pairs and its predictions on implied record pairs. TransClean iteratively modifies a matching through gradually removing false positive matches while removing as few true positive matches as possible. In each of these steps, the estimation of the Transitive Consistency is exclusively done through model evaluations and produces quantities that can be used as proxies of the amounts of true and false positives in the matching while not requiring any manual labeling, producing an estimate of the quality of the matching and indicating which record groups are likely to contain false positives. In our experiments, we compare combining TransClean with a naively trained pairwise matching model (DistilBERT) and with a state-of-the-art end-to-end matching method (CLER) and illustrate the flexibility of TransClean in being able to detect most of the false positives of either setup across a variety of datasets. Our experiments show that TransClean induces an average +24.42 F1 score improvement for entity matching in a multi-source setting when compared to traditional pair-wise matching algorithms.

Problem

• Detects false positives in multi-source entity matching
• Operates efficiently with noisy, unlabeled datasets
• Improves F1 score in entity matching tasks

Innovation

• Leverages Transitive Consistency for false positive detection
• Operates efficiently on large noisy multi-source datasets
• Requires minimal manual labeling for quality estimation