Multilingual End to End Entity Linking

Mikhail Plekhanov, Nora Kassner, Kashyap Popat, Louis Martin, Simone Merello +3 more
6/15/2023
cs.CL

No official code found. Best implementation: facebookresearch/BELA (50 stars, MIT, last commit 2024-09-10).

Abstract

Entity Linking is one of the most common Natural Language Processing tasks in practical applications, but so far efficient end-to-end solutions with multilingual coverage have been lacking, leading to complex model stacks. To fill this gap, we release and open source BELA, the first fully end-to-end multilingual entity linking model that efficiently detects and links entities in texts in any of 97 languages. We provide here a detailed description of the model and report BELA's performance on four entity linking datasets covering high- and low-resource languages.

View on arXivView PDF

Code Implementations(1)

facebookresearch/BELACommunity100%
MIT

Bi-encoder entity linking architecture

509Shell, PythonAug 31, 20212 years agoMIT

Similar Papers

Multilingual Entity Linking Using Dense Retrieval

85% match2024

ReFinED: An Efficient Zero-shot-capable Approach to End-to-End Entity Linking

84% match1 repo2502022

Instructed Language Models with Retrievers Are Powerful Entity Linkers

84% match1 repo62023

DBLPLink: An Entity Linker for the DBLP Scholarly Knowledge Graph

84% match2023

MERLIN: A Testbed for Multilingual Multimodal Entity Recognition and Linking

84% match1 repo32025

Contextual Augmentation for Entity Linking using Large Language Models

83% match2025

Cite this paper

@article{plekhanov2023multilingual,
  title  = {Multilingual End to End Entity Linking},
  author = {Mikhail Plekhanov and Nora Kassner and Kashyap Popat and Louis Martin and Simone Merello and Borislav Kozlovskii and Frédéric A. Dreyer and Nicola Cancedda},
  year   = {2023},
  eprint = {2306.08896},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2306.08896v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2306.08896.svg)](https://codeofpaper.com/paper/2306.08896)

Discussion