Domain-Specific Text Embedding Models for Entity Resolution
2026-08-17 • Information Retrieval
Information RetrievalArtificial IntelligenceMachine Learning
AI summaryⓘ
The authors studied how general text embedding models, which measure similarity between texts, struggle to tell apart records that look very similar but represent different real-world identities like businesses or people. They created a special dataset with tricky examples to test this. By fine-tuning the models using a specific training method called triplet training on domain-related data, they improved the models' ability to correctly identify whether two records refer to the same entity or not. Their work shows that customizing general models with targeted training can help in tasks like cleaning up duplicate records.
text embeddingsemantic similarityentity resolutiontriplet trainingfine-tuningidentity-sensitive retrievaldata quality managementinformation retrievalmargin-based evaluation
Authors
Khajesh Sapram, Srivardhani Raju, Kishore Konda
Abstract
General-purpose text embedding models are designed to capture semantic similarity but are not optimised for distinguishing entity records that represent the same real-world business or person. This limitation affects applications such as entity resolution and duplicate record retrieval, where small textual differences may either preserve or change identity. This paper investigates whether domain-specific triplet fine-tuning can adapt pretrained embedding models for identity-sensitive retrieval. A synthetic dataset of business and person records was created with identity-preserving variations and challenging non-matching examples. Two widely used embedding models were evaluated before and after fine-tuning using a margin-based similarity evaluation. The results show substantial improvements in separating true matches from highly similar non-matches, demonstrating that domain-specific triplet training can effectively reshape general-purpose embedding spaces for entity retrieval. These findings suggest that targeted fine-tuning provides a practical approach for improving embedding models in data quality management and information retrieval applications.