Structure-Guided Entity Resolution: Fine-Tuning LLMs for Robust Name Matching in Complex Linguistic Contexts
Published in Association for Computational Linguistics (ACL), 2026
Fine-tuning approach for LLM-based entity resolution targeting name-matching under noisy and linguistically diverse inputs (transliteration, abbreviation, code-switching). The structure-guided objective constrains the model to attend to script, phonetic, and morphological cues, producing higher-precision matches than vanilla instruction-tuned baselines on production-scale name pairs.
Recommended citation: Patil, N. et al. (2026). Structure-Guided Entity Resolution: Fine-Tuning LLMs for Robust Name Matching in Complex Linguistic Contexts. Association for Computational Linguistics.
Download Paper