LELA: An End-to-end LLM-based Entity Linking Framework with Zero-shot Domain Adaptation

2026-05-26

AI Summary

This work proposes an end-to-end entity linking framework based on large language models (LLMs) that overcomes the limited generalizability of existing approaches, which often rely on domain-specific knowledge bases. The method uniquely integrates zero-shot named entity recognition with a modular, domain-agnostic entity disambiguation component, enabling cross-domain deployment without fine-tuning. Designed with a modular architecture and encapsulated through engineered Python implementation, the framework demonstrates strong performance and robustness across diverse settings. An accompanying interactive demonstration system is also provided to facilitate real-time validation and usability.

Abstract

Entity linking is a key component of many downstream NLP systems, yet existing approaches are often tied to the specific target knowledge bases and domains, limiting their real world application. In this paper, we extend LELA, a modular and domain-agnostic LLM-based entity disambiguation method, into a practical Python library that integrates zero-shot Named Entity Recognition (NER) -thereby providing a complete end-toend pipeline for entity-linking in real-world usage. We provide experimental results validating LELA's performance and robustness across diverse entity linking settings. In our demo, users can play with the system on their own input texts.

Problem

• Entity Linking
• Domain Adaptation
• Knowledge Base
• Named Entity Recognition
• LLM-based

Innovation

• entity linking
• large language model
• zero-shot learning
• domain adaptation
• named entity recognition