F-COREF: Fast, Accurate and Easy to Use Coreference Resolution

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

27 Scopus citations

Abstract

We introduce fastcoref, a python package for fast, accurate, and easy-to-use English coreference resolution. The package is pip-installable, and allows two modes: an accurate mode based on the LINGMESS architecture, providing state-of-the-art coreference accuracy, and a substantially faster model, F-COREF, which is the focus of this work. F-COREF allows to process 2.8K OntoNotes documents in 25 seconds on a V100 GPU (compared to 6 minutes for the LINGMESS model, and to 12 minutes of the popular AllenNLP coreference model) with only a modest drop in accuracy. The fast speed is achieved through a combination of distillation of a compact model from the LingMess model, and an efficient batching implementation using a technique we call leftover batching.

Original languageEnglish
Title of host publicationSystem Demonstrations
EditorsWray Buntine, Maria Liakata
PublisherAssociation for Computational Linguistics (ACL)
Pages48-56
Number of pages9
ISBN (Electronic)9781955917551
DOIs
StatePublished - 2022
Event2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing, AACL-IJCNLP 2022 - Virtual, Online
Duration: 20 Nov 202223 Nov 2022

Publication series

NameProceedings of the 2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing: Long Paper, AACL-IJCNLP 2022
Volume4

Conference

Conference2nd Conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint Conference on Natural Language Processing, AACL-IJCNLP 2022
CityVirtual, Online
Period20/11/2223/11/22

Bibliographical note

Publisher Copyright:
©2022 Association for Computational Linguistics.

Fingerprint

Dive into the research topics of 'F-COREF: Fast, Accurate and Easy to Use Coreference Resolution'. Together they form a unique fingerprint.

Cite this