Towards Comprehensive Detection of Chinese Harmful Memes
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909334262251520 |
|---|---|
| author | Lu, Junyu Xu, Bo Zhang, Xiaokun Wang, Hongbo Zhu, Haohao Zhang, Dongyu Yang, Liang Lin, Hongfei |
| author_facet | Lu, Junyu Xu, Bo Zhang, Xiaokun Wang, Hongbo Zhu, Haohao Zhang, Dongyu Yang, Liang Lin, Hongfei |
| contents | This paper has been accepted in the NeurIPS 2024 D & B Track. Harmful memes have proliferated on the Chinese Internet, while research on detecting Chinese harmful memes significantly lags behind due to the absence of reliable datasets and effective detectors. To this end, we focus on the comprehensive detection of Chinese harmful memes. We construct ToxiCN MM, the first Chinese harmful meme dataset, which consists of 12,000 samples with fine-grained annotations for various meme types. Additionally, we propose a baseline detector, Multimodal Knowledge Enhancement (MKE), incorporating contextual information of meme content generated by the LLM to enhance the understanding of Chinese memes. During the evaluation phase, we conduct extensive quantitative experiments and qualitative analyses on multiple baselines, including LLMs and our MKE. The experimental results indicate that detecting Chinese harmful memes is challenging for existing models while demonstrating the effectiveness of MKE. The resources for this paper are available at https://github.com/DUT-lujunyu/ToxiCN_MM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_02378 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Towards Comprehensive Detection of Chinese Harmful Memes Lu, Junyu Xu, Bo Zhang, Xiaokun Wang, Hongbo Zhu, Haohao Zhang, Dongyu Yang, Liang Lin, Hongfei Computation and Language Artificial Intelligence This paper has been accepted in the NeurIPS 2024 D & B Track. Harmful memes have proliferated on the Chinese Internet, while research on detecting Chinese harmful memes significantly lags behind due to the absence of reliable datasets and effective detectors. To this end, we focus on the comprehensive detection of Chinese harmful memes. We construct ToxiCN MM, the first Chinese harmful meme dataset, which consists of 12,000 samples with fine-grained annotations for various meme types. Additionally, we propose a baseline detector, Multimodal Knowledge Enhancement (MKE), incorporating contextual information of meme content generated by the LLM to enhance the understanding of Chinese memes. During the evaluation phase, we conduct extensive quantitative experiments and qualitative analyses on multiple baselines, including LLMs and our MKE. The experimental results indicate that detecting Chinese harmful memes is challenging for existing models while demonstrating the effectiveness of MKE. The resources for this paper are available at https://github.com/DUT-lujunyu/ToxiCN_MM. |
| title | Towards Comprehensive Detection of Chinese Harmful Memes |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2410.02378 |