Character-level HyperNetworks for Hate Speech Detection

作者：

Highlights：

• Present compact character-based HyperNetworks for hate speech detection.

• These networks are trained using large-scale amounts of generated data.

• Performance is competitive and sometimes exceeds large pretrained language models.

• The proposed compact solution can fit end devices.

摘要

•Present compact character-based HyperNetworks for hate speech detection.•These networks are trained using large-scale amounts of generated data.•Performance is competitive and sometimes exceeds large pretrained language models.•The proposed compact solution can fit end devices.

论文关键词：Hate speech detection,Neural networks,Text generation

论文评审过程：Received 7 July 2021, Revised 4 April 2022, Accepted 9 May 2022, Available online 17 May 2022, Version of Record 31 May 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.117571