FedAds
收藏资源简介:
FedAds数据集是由阿里巴巴集团基于其广告平台收集的大规模真实世界数据集,专门用于隐私保护的CVR估计与垂直联邦学习研究。该数据集包含1130万条样本,涵盖用户点击事件及其后续行为,如购买行为等。数据集的创建旨在通过系统化的评估,促进垂直联邦学习算法的发展,特别是在提高模型效果和保护隐私方面的应用。FedAds数据集的应用领域主要集中在在线广告和推荐系统中,旨在通过联邦学习技术,不交换原始数据的情况下,结合多方优势,提高CVR估计的准确性和用户数据隐私的保护。
The FedAds dataset is a large-scale real-world dataset collected by Alibaba Group based on its advertising platform, specifically tailored for privacy-preserving CVR estimation and vertical federated learning research. It contains 11.3 million samples, covering user click events and their subsequent behaviors such as purchase actions. The dataset was developed to facilitate the advancement of vertical federated learning algorithms via systematic evaluation, particularly their applications in improving model performance and safeguarding privacy. The FedAds dataset is mainly applied in online advertising and recommendation systems, where it aims to integrate the advantages of multiple parties through federated learning technologies without exchanging raw user data, thereby enhancing the accuracy of CVR estimation and protecting user data privacy.

- 1FedAds: A Benchmark for Privacy-Preserving CVR Estimation with Vertical Federated Learning阿里巴巴集团 · 2023年



