RECALL-MM
收藏资源简介:
RECALL-MM数据集是由加州大学伯克利分校和Autodesk Research共同创建的多模态数据集,包含6874条美国消费品安全委员会(CPSC)的召回记录,时间跨度为2000年至2024年。数据集通过大型语言模型(LLM)对召回描述和产品名称进行了增强,并提供了产品的文本描述、图像以及召回相关的分类信息。该数据集旨在为风险分析提供数据驱动的支持,通过历史信息增强产品安全设计,并可用于预测潜在的产品风险,指导更安全的设计决策。
The RECALL-MM dataset is a multimodal dataset jointly created by the University of California, Berkeley and Autodesk Research. It contains 6,874 recall records from the U.S. Consumer Product Safety Commission (CPSC) spanning the period from 2000 to 2024. The dataset has been enhanced for recall descriptions and product names using Large Language Models (LLMs), and provides textual product descriptions, images, as well as classification information related to product recalls. This dataset aims to provide data-driven support for risk analysis, enhance product safety design through historical information, and can be used to predict potential product risks to guide safer design decisions.




