FakeVV
收藏资源简介:
FakeVV是迄今为止最大的、最全面标注的新闻领域视频 misinformation 数据集。它具有高质量、生成的视频字幕,专门设计用于支持 misinformation 推理任务。数据集内容丰富,包括视频、标题、字幕和元数据等,数据来源为 BBC News、Guardian News、CNN 和 The New York Times 等官方新闻渠道。数据集创建过程分为数据收集、数据预处理和 misinformation 数据构建三个阶段,旨在解决当前 misinformation 检测领域中存在的规模、多样性和可解释性等问题。
FakeVV is the largest and most comprehensively annotated video misinformation dataset in the news domain to date. It features high-quality generated video captions, and is specifically designed to support misinformation reasoning tasks. The dataset has rich content including videos, headlines, captions, metadata and other related materials, with data sourced from official news outlets such as BBC News, Guardian News, CNN and The New York Times. The construction of the dataset is divided into three stages: data collection, data preprocessing and misinformation data construction, aiming to address the core challenges of scale, diversity and interpretability in the current misinformation detection field.
Fact-R1数据集概述
基本信息
- 数据集名称:Fact-R1
- 当前状态:Under Review(审核中)
补充说明
- 该数据集目前未提供详细描述信息
- 无其他可用元数据或使用说明

- 1Fact-R1: Towards Explainable Video Misinformation Detection with Deep Reasoning中国科学技术大学, 腾讯QQ, 上海创新研究院, 复旦大学 · 2025年



