SemEval-2022 Task 5: MAMI - Multimedia Automatic Misogyny Identification

该数据集旨在通过考虑文本和视觉内容，识别带有性别歧视的迷因图。在此挑战中，研究者需预测一个迷因图是否具有性别歧视倾向，并识别性别歧视的不同子类别。该数据集包含了从图像中提取的叠加文本，并分为两个子任务：任务A（预测性别歧视）和任务B（识别性别歧视的子类别）。数据集规模包括10,000个训练样本和1,000个测试样本，旨在实现多媒体自动性别歧视识别任务。

This dataset is designed to identify sexist memes by taking into account both their textual and visual content. In this challenge, participants are tasked with two objectives: predicting whether a given meme is sexist, and identifying distinct subcategories of sexism within the content. The dataset includes overlay text extracted from the meme images, and is split into two subtasks: Task A (Sexism Prediction) and Task B (Sexism Subcategory Identification). Comprising 10,000 training samples and 1,000 test samples, this dataset supports research on automated multimedia sexism recognition.