May 8, 2024Open Access

LinkNER: Linking Local Named Entity Recognition Models to Large Language Models using Uncertainty

View Full Paper

Ask AI

Bookmark

Authors

Discussion

LinkNER enhances named entity recognition performance in web applications, indicating improved usability and reliability.

LinkNER improves named entity recognition task performance, especially on unseen entities and diverse datasets.
Enhanced robustness is shown, surpassing state-of-the-art models through a novel uncertainty linking method.
Assessment using both fine-tuned named entity recognition models and large language models contributes to overall reliability in performance metrics and web applications optimization factors . The proposed framework combines strengths of small fine-tuned models with the knowledge of large language models, streamlining the entity recognition process.

Zhang et al. (2024) studied this question.

View Full Paper

Ask AI

Bookmark