arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2003.08082cs.LGcs.CVstat.ML
收藏

真实世界数据分布的联邦视觉分类

Federated Visual Classification with Real-World Data Distribution

MIT CSAIL(麻省理工学院计算机科学与人工智能实验室);Google Research(谷歌研究院)

Tzu-Ming Harry Hsu, Hang Qi, Matthew Brown

摘要:

联邦学习使视觉模型能够在设备上训练,为用户隐私带来优势(数据永远不需要离开设备),但在数据多样性和质量方面存在挑战。虽然数据中心中的典型模型是使用独立同分布(IID)的数据进行训练的,但源数据通常远离IID。此外,不同数量的数据通常在每个设备处可用(不平衡)。在这项工作中,我们描述了这些真实世界的数据分布对分布式学习的影响,使用标准的联邦平均(FedAvg)算法作为基准。为此,我们引入了两个新的用于物种和地标分类的大规模数据集,其中每个用户数据分割都模拟了真实世界的边缘学习场景。我们还开发了两种新算法(FedVC,FedIR),可以智能地对客户端池进行重新采样和重新加权,从而大大提高了训练的准确性和稳定性。这些数据集可在网上查阅。

英文摘要:

Federated Learning enables visual models to be trained on-device, bringing advantages for user privacy (data need never leave the device), but challenges in terms of data diversity and quality. Whilst typical models in the datacenter are trained using data that are independent and identically distributed (IID), data at source are typically far from IID. Furthermore, differing quantities of data are typically available at each device (imbalance). In this work, we characterize the effect these real-world data distributions have on distributed learning, using as a benchmark the standard Federated Averaging (FedAvg) algorithm. To do so, we introduce two new large-scale datasets for species and landmark classification, with realistic per-user data splits that simulate real-world edge learning scenarios. We also develop two new algorithms (FedVC, FedIR) that intelligently resample and reweight over the client pool, bringing large improvements in accuracy and stability in training. The datasets are made available online.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑