MemeXplain

# ArMeme Explanation Dataset

This dataset contains Arabic meme images with explanations for propaganda detection.

## Dataset Description

The ArMeme Explanation dataset provides detailed explanations for Arabic memes, helping to understand whether they contain propaganda content.

## Dataset Structure

Each entry in the dataset contains:
- `id`: Unique identifier for the meme
- `text`: Arabic text extracted from the meme
- `img_path`: Path to the meme image
- `class_label`: Classification label (`propaganda` or `not-propaganda`)
- `explanation_en`: English explanation of the classification
- `explanation_ar`: Arabic explanation of the classification

## Dataset Splits

- **Train**: 4,007 examples
- **Dev**: 580 examples
- **Test**: 1,134 examples

## Citation

If you use this dataset, please cite:

```bibtex
@misc{kmainasi2025memeintelexplainabledetectionpropagandistic,
title={MemeIntel: Explainable Detection of Propagandistic and Hateful Memes},
author={Mohamed Bayan Kmainasi and Abul Hasnat and Md Arid Hasan and Ali Ezzat Shahroor and Firoj Alam},
year={2025},
eprint={2502.16612},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2502.16612},
}
```

Please also cite the original ArMeme dataset:

```bibtex
@inproceedings{alam-etal-2024-armeme,
title = "{A}r{M}eme: Propagandistic Content in {A}rabic Memes",
author = "Alam, Firoj and
Hasnat, Abul and
Ahmad, Fatema and
Hasan, Md. Arid and
Hasanain, Maram",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = "2024",
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.1173/",
doi = "10.18653/v1/2024.emnlp-main.1173",
pages = "21071--21090",
}
```

## License

This dataset is licensed under Creative Commons License (CC BY 4.0). Please refer to the original ArMeme dataset license for details.

# ArMeme 解释数据集

本数据集收录了用于宣传内容检测的阿拉伯语表情包（meme）及其配套解释文本。

## 数据集说明

ArMeme 解释数据集为阿拉伯语表情包提供详细解读，助力研究者判断其是否包含宣传类内容。

## 数据集结构

本数据集的每条数据包含以下字段：
- `id`：表情包的唯一标识符
- `text`：从表情包中提取的阿拉伯语文本
- `img_path`：表情包图像的存储路径
- `class_label`：分类标签，取值为`propaganda`（宣传类）或`not-propaganda`（非宣传类）
- `explanation_en`：该分类结果的英文解释
- `explanation_ar`：该分类结果的阿拉伯语解释

## 数据集划分

- **训练集（Train）**：4007条样本
- **验证集（Dev）**：580条样本
- **测试集（Test）**：1134条样本

## 引用说明

若使用本数据集，请引用以下文献：
bibtex
@misc{kmainasi2025memeintelexplainabledetectionpropagandistic,
title={MemeIntel: Explainable Detection of Propagandistic and Hateful Memes},
author={Mohamed Bayan Kmainasi and Abul Hasnat and Md Arid Hasan and Ali Ezzat Shahroor and Firoj Alam},
year={2025},
eprint={2502.16612},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/2502.16612},
}

同时请一并引用原始ArMeme数据集：
bibtex
@inproceedings{alam-etal-2024-armeme,
title = "{A}r{M}eme: Propagandistic Content in {A}rabic Memes",
author = "Alam, Firoj and
Hasnat, Abul and
Ahmad, Fatema and
Hasan, Md. Arid and
Hasanain, Maram",
editor = "Al-Onaizan, Yaser and
Bansal, Mohit and
Chen, Yun-Nung",
booktitle = "Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing",
month = nov,
year = 2024,
address = "Miami, Florida, USA",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.emnlp-main.1173/",
doi = "10.18653/v1/2024.emnlp-main.1173",
pages = "21071--21090",
}

## 授权协议

本数据集采用知识共享署名4.0国际许可协议（Creative Commons License (CC BY 4.0)）进行授权，详细许可条款请参阅原始ArMeme数据集的授权协议。