# Entity Linking via Dual and Cross-Attention Encoders

arXiv: https://arxiv.org/abs/2004.03555 · HTML: https://mlsanity.com/paper/2004.03555/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 7 Apr 2020.

## Built on by (most cited first, 4 of 4)

- [QUEACO: Borrowing Treasures from Weakly-labeled Behavior Data for Query Attribute Value Extraction](https://mlsanity.com/paper/2108.08468/) (2021)
- [MOLEMAN: Mention-Only Linking of Entities with a Mention Annotation Network](https://mlsanity.com/paper/2106.07352/) (2021)
- [Improving Candidate Retrieval with Entity Profile Generation for Wikidata Entity Linking](https://mlsanity.com/paper/2202.13404/) (2022)
- [Personalized Entity Resolution with Dynamic Heterogeneous Knowledge Graph Representations](https://mlsanity.com/paper/2104.02667/) (2021)

## Abstract (arXiv)

Entity Linking has two main open areas of research: 1) generate candidate entities without using alias tables and 2) generate more contextual representations for both mentions and entities. Recently, a solution has been proposed for the former as a dual-encoder entity retrieval system (Gillick et al., 2019) that learns mention and entity representations in the same space, and performs linking by selecting the nearest entity to the mention in this space. In this work, we use this retrieval system solely for generating candidate entities. We then rerank the entities by using a cross-attention encoder over the target mention and each of the candidate entities. Whereas a dual encoder approach forces all information to be contained in the small, fixed set of vector dimensions used to represent mentions and entities, a crossattention model allows for the use of detailed information (read: features) from the entirety of each <mention, context, candidate entity> tuple. We experiment with features used in the reranker including different ways of incorporating document-level context. We achieve state-of-the-art results on TACKBP-2010 dataset, with 92.05% accuracy. Furthermore, we show how the rescoring model generalizes well when trained on the larger CoNLL-2003 dataset and evaluated on TACKBP-2010.