Edinburgh at SemEval-2022 Task 1: Jointly Fishing for Word Embeddings and Definitions

Pinzhen Chen, Zheng Zhao

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

This paper presents a winning submission to the SemEval 2022 Task 1 on two sub-tasks: reverse dictionary and definition modelling. We leverage a recently proposed unified model with multi-task training. It utilizes data symmetrically and learns to tackle both tracks concurrently. Analysis shows that our system performs consistently on diverse languages, and works the best with sgns embeddings. Yet, char and electra carry intriguing properties. The two tracks' best results are always in differing subsets grouped by linguistic annotations. In this task, the quality of definition generation lags behind, and BLEU scores might be misleading.
Original languageEnglish
Title of host publicationProceedings of the 16th International Workshop on Semantic Evaluation (SemEval-2022)
EditorsGuy Emerson, Natalie Schluter, Gabriel Stanovsky, Ritesh Kumar, Alexis Palmer, Nathan Schneider, Siddharth Singh, Shyam Ratan
Place of PublicationSeattle, United States
PublisherAssociation for Computational Linguistics
Pages75-81
Number of pages7
DOIs
Publication statusPublished - 1 Jul 2022
EventThe 16th International Workshop on Semantic Evaluation 2022
-
Duration: 14 Jul 202215 Jul 2022
Conference number: 16
https://semeval.github.io/SemEval2022/

Workshop

WorkshopThe 16th International Workshop on Semantic Evaluation 2022
Abbreviated titleSemEval 2022
Period14/07/2215/07/22
Internet address

Fingerprint

Dive into the research topics of 'Edinburgh at SemEval-2022 Task 1: Jointly Fishing for Word Embeddings and Definitions'. Together they form a unique fingerprint.

Cite this