TWIGMA
收藏资源简介:
TWIGMA是由斯坦福大学开发的包含80万张AI生成图像的数据集,涵盖了2021年至2023年期间从Twitter收集的图像及其元数据。该数据集不仅包括图像本身,还包含如推文文本、创建日期和点赞数等元数据,支持对AI生成图像的内容和主题变化进行时间序列分析。TWIGMA的创建旨在探索AI生成图像的多样性和用户偏好,特别是在社交媒体平台上的表现,为研究AI生成图像的社会影响和人类与AI交互提供了独特的数据资源。
TWIGMA is a dataset containing 800,000 AI-generated images developed by Stanford University, covering images and their associated metadata collected from Twitter between 2021 and 2023. Beyond the images themselves, the dataset also includes metadata such as tweet text, creation date, and like counts, enabling time-series analysis of the content and thematic shifts of AI-generated images. TWIGMA was created to explore the diversity of AI-generated images and user preferences, especially their performance on social media platforms, providing a unique data resource for researching the societal impact of AI-generated images and human-AI interaction.

- 1TWIGMA: A dataset of AI-Generated Images with Metadata From Twitter斯坦福大学 · 2023年



