Network embeddings from distributional thesauri for improving static word representations

作者：

Highlights：

• Distributional thesaurus embedding helps to improve the SOTA static word embeddings.

• Proposed embedding boosts performance of various intrinsic and extrinsic NLP tasks.

• The code and embeddings used in this study will be made publicly available.

摘要

•Distributional thesaurus embedding helps to improve the SOTA static word embeddings.•Proposed embedding boosts performance of various intrinsic and extrinsic NLP tasks.•The code and embeddings used in this study will be made publicly available.

论文关键词：Word representation,Network embedding,Distributional thesaurus,Word level semantics,Distributional hypothesis

论文评审过程：Received 2 May 2020, Revised 16 February 2021, Accepted 2 September 2021, Available online 14 September 2021, Version of Record 21 September 2021.

论文官网地址：https://doi.org/10.1016/j.eswa.2021.115868