Empirical Evaluation of Pretraining Strategies for Supervised Entity Linking
This work addresses entity linking for natural language processing applications, presenting incremental improvements through empirical evaluation of pretraining strategies.
The paper tackles the problem of entity linking by combining a Transformer architecture with large-scale pretraining from Wikipedia links, achieving state-of-the-art results of 96.7% on CoNLL and 94.9% on TAC-KBP datasets.
In this work, we present an entity linking model which combines a Transformer architecture with large scale pretraining from Wikipedia links. Our model achieves the state-of-the-art on two commonly used entity linking datasets: 96.7% on CoNLL and 94.9% on TAC-KBP. We present detailed analyses to understand what design choices are important for entity linking, including choices of negative entity candidates, Transformer architecture, and input perturbations. Lastly, we present promising results on more challenging settings such as end-to-end entity linking and entity linking without in-domain training data.