Irony detection via sentiment-based transfer learning

作者：

Highlights：

• Take advantage of the readily available sentiment resources to identify implicit incongruity for irony detection.

• Transferring deep sentiment features to a neural attention model is an effective approach to extract patterns of implicit incongruity embedded in ironic texts.

• Evaluate irony detection models using human-annotated and automatic hashtag-labeled datasets separately.

摘要

•Take advantage of the readily available sentiment resources to identify implicit incongruity for irony detection.•Transferring deep sentiment features to a neural attention model is an effective approach to extract patterns of implicit incongruity embedded in ironic texts.•Evaluate irony detection models using human-annotated and automatic hashtag-labeled datasets separately.

论文关键词：Irony detection,Transfer learning

论文评审过程：Received 6 September 2018, Revised 18 April 2019, Accepted 19 April 2019, Available online 16 May 2019, Version of Record 16 May 2019.

论文官网地址：https://doi.org/10.1016/j.ipm.2019.04.006