Offensive, aggressive, and hate speech analysis: From data-centric to human-centered approach

作者:

Highlights:

• A list of requirements for new human-centered content analyses.

• 3 perspectives for subjective content perception that affect annotation procedures.

• New methods for offensive content recognition that respect full personalization.

• Reasoning process based on personal behavior implementing human-centered approaches.

• Multimodal transformer-based human-centered data representations.

• Combining text, metadata, annotator demographics, and annotator behavior.

• Evidence that human-centered classification approaches outperform classic methods.

摘要

•A list of requirements for new human-centered content analyses.•3 perspectives for subjective content perception that affect annotation procedures.•New methods for offensive content recognition that respect full personalization.•Reasoning process based on personal behavior implementing human-centered approaches.•Multimodal transformer-based human-centered data representations.•Combining text, metadata, annotator demographics, and annotator behavior.•Evidence that human-centered classification approaches outperform classic methods.

论文关键词:Hate speech,Offensive content,Human-centered NLP,Multimodal deep learning,Personalization,Subjective content perception,Annotator agreement

论文评审过程:Received 23 December 2020, Revised 11 March 2021, Accepted 14 May 2021, Available online 3 June 2021, Version of Record 3 June 2021.

论文官网地址:https://doi.org/10.1016/j.ipm.2021.102643