This repository contains official implementation of the paper ViHOS: Vietnamese Hate and Offensive Spans Detection accepted at the EACL 2023 Main Conference.

Disclaimer: This project contains real comments that could be considered profane, offensive, or abusive.

The rise in hateful and offensive language directed at other users is one of the adverse side effects of the increased use of social networking platforms. This could make it difficult for human moderators to review tagged comments filtered by classification systems.

To help address this issue, we present the ViHOS (Vietnamese Hate and Offensive Spans) dataset, the first human-annotated corpus containing 26k spans on 11k online comments.

Our goal is to create a dataset that contains comprehensive hate and offensive thoughts, meanings, or opinions within the comments rather than just a lexicon of hate and offensive terms.

We also provide definitions of hateful and offensive spans in Vietnamese comments as well as detailed annotation guidelines. Futhermore, our solutions to deal with nine different online foul linguistic phenomena are also provided in the paper (e.g. Teencodes; Metaphors, metonymies; Hyponyms; Puns...).

We hope that this dataset will be useful for researchers and practitioners in the field of hate speech detection in general and hate spans detection in particular.

ViHOS contains 26,476 human-annotated spans on 11,056 comments (5,360 comments have hate and offensive spans, and 5,696 comments do not)

It is splitted into train, dev, and test set with following information:

1. Train set: 8,844 comments
2. Dev set: 1,106 comments
3. Test set: 1,106 comments

See README.md in the dataset folder for more details.

Table 1. ViHOS statistics. Vocabularies size and comments length are calculated at the syllable level. In which, Ha/Off? stands for a hate (Ha) or offensive (Off).

Table 2. Spans quantity and length statistics.

NOTE: Our dataset has equal number of span and non-span (clean) comments because:

1. We aim to detect the hate and offensive spans directly in online comments
2. With an equal number of span and non-span comments helps models not be biased towards any type.

Table 3. Experimental results on Full Data versus Without additional clean comments.

If you use our dataset, codes or analyses, please cite our paper:

@article{hoang2023vihos,
  title={ViHOS: Hate Speech Spans Detection for Vietnamese},
  author={Hoang, Phu Gia and Luu, Canh Duc and Tran, Khanh Quoc and Van Nguyen, Kiet and Nguyen, Ngan Luu-Thuy},
  journal={arXiv preprint arXiv:2301.10186},
  year={2023}
}