Similarly, Yu et al. (2020) build a federated multi- task learning framework for smart home IOT to automatically learn users’ behavior patterns, which could effectively detect physical hazards. Furthermore, Liu, Wang, Liu, and Xu (2020) proposed a data fusion approach based on FL for robots...
数据分割 在跨孤岛联邦学习设置下,数据可以按样本分割——横向联邦学习(Horizontal Federated Learning),可以按特征分割——纵向联邦学习(Vertical Federated Learning)。 按特征分割和按样本分割使用的训练体系有很大不同,它可能涉及也可能不涉及中央服务器(取决于是中心化学习或去中心化学习),并基于训练算法的特性,客户...
Communication costs(通信成本):通过设置了在联邦学习训练轮次期间选择客户端更新的预定义规则,有助于通过消除贡献最少的联邦学习客户端更新来最小化通信成本(论文链接:Client Selection for Federated Learning with Heterogeneous Resources in Mobile Edge)。 联邦蒸馏(Federated Distillation, FD)和联合增强(Federated Augm...
写不出论文了,兴趣使然地开一个论文笔记系列,不知道毕业的时候能写到多少篇(苦笑)。论文名称'Clustered federated learning: Model-Agnostic distributed multi-Task optimization under privacy constrain
Jiayi Chen and Aidong Zhang. 2022. FedMSplit: Correlation-Adaptive Federated Multi-Task Learning across Multimodal Split Networks. In Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD '22). Association for Computing Machinery, New York, NY, USA, 87–96....
技术标签:联邦学习论文笔记机器学习安全 Byzantine Resilient Distributed Multi-Task Learning Recently I got an intern chance and the topic is mainly in FL area. Thus it’s suitable to record the learning process and upload some notes to... ...
当以“federated personal”为关键词,在 dblp 上进行文献检索的时候,会发现个性化联邦学习(PFL, Personalized Federated Learning)的研究在2021年的有明显增加。 经小鱼的查询,至22年4月16日针对个性化联邦学习的综述文章仅两篇[1][2],其中...
阅读全文 赞同1 添加评论 分享收藏 [ICML 2021] Fair and Robust Federated Learning Through Personalization 我是一颗小蘑菇 与天斗,其乐无穷 我是一颗小蘑菇: 新鲜出炉的ICML2021的论文,趁热读了(因为后天开组会hhh) 原来的题目Federated Multi-Task Learning for Competing Constraint…阅读...
learning can lead to privacy leakage of the training data. To address the aforementioned challenges, we propose a privacy-preserving Byzantine-robust federated learning scheme (PBFL) which takes both the robustness of federated learning and the privacy of the workers into account. PBFL is ...
VS 哪些客户会买保险），但它们仍然可以加入同一个FL system。比如可以运用multi-task learning的思想来...