使用多模态大语言模型检测新加坡语境下带有社会偏见的冒犯性模因
Detecting Offensive Memes with Social Biases in Singapore Context Using Multimodal Large Language Models
AI总结:
本文针对新加坡多元文化语境下冒犯性模因检测难题,构建了112K规模的GPT-4V标注数据集,并提出结合OCR、翻译与70亿参数级VLM的检测流水线,在测试集上取得80.62%准确率和0.8192 AUROC。
摘要:
传统的在线内容审核系统难以对模因这类现代多模态传播方式进行分类,因为模因是一种高度微妙且信息密集的媒介。在新加坡这样文化多元的社会中,这一任务尤其困难,因为需要使用低资源语言,并且解读在线内容需要大量本地语境知识。我们整理了一个包含112K个模因的大型数据集,由GPT-4V进行标注,用于微调视觉语言模型(VLM),以对新加坡语境下的冒犯性模因进行分类。我们展示了微调后的VLM在该数据集上的有效性,并提出了一条包含OCR、翻译和一个70亿参数级VLM的流水线。我们的方案在留出测试集上达到了80.62%的准确率和0.8192的AUROC,能够极大地辅助人工进行在线内容审核。数据集、代码和模型权重已在 https://github.com/aliencaocao/vlm-for-memes-aisg 开源。
英文摘要:
Traditional online content moderation systems struggle to classify modern multimodal means of communication, such as memes, a highly nuanced and information-dense medium. This task is especially hard in a culturally diverse society like Singapore, where low-resource languages are used and extensive knowledge on local context is needed to interpret online content. We curate a large collection of 112K memes labeled by GPT-4V for fine-tuning a VLM to classify offensive memes in Singapore context. We show the effectiveness of fine-tuned VLMs on our dataset, and propose a pipeline containing OCR, translation and a 7-billion parameter-class VLM. Our solutions reach 80.62% accuracy and 0.8192 AUROC on a held-out test set, and can greatly aid human in moderating online contents. The dataset, code, and model weights have been open-sourced at https://github.com/aliencaocao/vlm-for-memes-aisg.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。