首个中文有害表情包数据集，提升对中文网络毒梗的识别能力
Towards Comprehensive Detection of Chinese Harmful Memes
- 构建1.2万样本中文有害表情包数据集，标注细致
- 提出多模态知识增强模型，结合大模型上下文理解
- 实验证明现有模型难检测中文毒梗，新方法有效
本文被NeurIPS 2024 D&B Track接收。中文网络中有害表情包泛滥，但因缺乏可靠数据集与有效检测器，相关研究严重滞后。为此，我们聚焦中文有害表情包的全面检测，构建了首个中文有害表情包数据集ToxiCN MM，包含12,000个样本，并对多种表情包类型进行细粒度标注。同时，提出基线检测器多模态知识增强（MKE），利用大语言模型生成的表情包内容上下文信息，提升对中文表情包的理解。在评估阶段，我们在多个基线模型（包括大语言模型和MKE）上开展广泛定量实验与定性分析。结果表明，现有模型在检测中文有害表情包方面表现不佳，而MKE展现出显著有效性。相关资源已开源：https://github.com/DUT-lujunyu/ToxiCN_MM。
原文摘要 · Abstract (English)
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.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。