Introducing high correlation and high quality instances for few-shot entity linking

Published: 01 Jan 2025, Last Modified: 05 Mar 2025Neural Networks 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•First prioritize the correlation of instances to the target domain.•Adversarial extraction for high correlation instances from high quality data.•Curriculum learning variant to mitigate low-correlation instances’ negative effect.•The micro accuracy on the few-shot entity linking dataset improved 19.47%.
Loading