Chinese named-entity recognition via self-attention mechanism and position-aware influence propagation embedding

作者：

Highlights：

• Globally modeling the entire sequence by the self-attention mechanism.

• The surrounding characters spread their influence to the represented character.

• Applying the Gaussian kernel for each character; alleviate the problem of the lack of word boundary information for the character-based methods.

• Carry on more parallelization to improve the inference efficiency.

摘要

•Globally modeling the entire sequence by the self-attention mechanism.•The surrounding characters spread their influence to the represented character.•Applying the Gaussian kernel for each character; alleviate the problem of the lack of word boundary information for the character-based methods.•Carry on more parallelization to improve the inference efficiency.

论文关键词：Chinese named-entity recognition,Self-attention,Position-aware influence propagation,Gaussian kernel

论文评审过程：Received 25 October 2020, Revised 8 October 2021, Accepted 10 January 2022, Available online 21 January 2022, Version of Record 12 February 2022.

论文官网地址：https://doi.org/10.1016/j.datak.2022.101983