Sentence modeling via multiple word embeddings and multi-level comparison for semantic textual similarity

作者：

Highlights：

• Encoding sentence via multiple pre-trained word embeddings.

• Evaluating sentence pairs via multi-levels comparison.

• The approach achieves strong performances on semantic textual similarity tasks.

• The approach does not rely on linguistic resources.

摘要

•Encoding sentence via multiple pre-trained word embeddings.•Evaluating sentence pairs via multi-levels comparison.•The approach achieves strong performances on semantic textual similarity tasks.•The approach does not rely on linguistic resources.

论文关键词：Multiple word embeddings,Sentence embedding,Semantic,Similarity,Multi-level comparison

论文评审过程：Received 14 February 2019, Revised 17 June 2019, Accepted 21 July 2019, Available online 9 August 2019, Version of Record 9 August 2019.

论文官网地址：https://doi.org/10.1016/j.ipm.2019.102090