Client Gradient

Client gradients, the updates calculated by individual devices in federated learning (FL), are central to training shared models on decentralized data without compromising privacy. Current research focuses on improving the efficiency and robustness of FL by addressing challenges like client heterogeneity (e.g., using weighted averaging or gradient masking to mitigate the influence of outliers), privacy vulnerabilities (e.g., developing defenses against gradient inversion attacks), and optimizing client selection strategies (e.g., prioritizing clients with aligned objectives or employing diverse sampling techniques). These advancements are crucial for enabling secure and effective FL across diverse applications, particularly in sensitive domains like healthcare and finance.

Papers

October 6, 2023

Utilizing Free Clients in Federated Learning for Focused Model Enhancement
Aditya Narayan Ravi, Ilan Shomorony
Faster Convergence Model Improvement Priority User Adaptive Federated Learning Client Gradient

September 14, 2023

Client-side Gradient Inversion Against Federated Learning from Poisoning
Jiaheng Wei, Yanjun Zhang, Leo Yu Zhang, Chao Chen, Shirui Pan, Kok-Leong Ong, Jun Zhang, Yang Xiang
Poisoning Attack Gradient Inversion Attack Byzantine Robust Aggregation Client Gradient

June 14, 2023

Provably Personalized and Robust Federated Learning
Mariel Werner, Lie He, Michael Jordan, Martin Jaggi, Sai Praneeth Karimireddy
Personalized Subject Optimal Convergence Rate Client Gradient

March 21, 2023

LOKI: Large-scale Data Reconstruction Attack against Federated Learning through Model Manipulation
Joshua C. Zhao, Atul Sharma, Ahmed Roushdy Elkordy, Yahya H. Ezzeldin, Salman Avestimehr, Saurabh Bagchi
Federated Learning Secure Aggregation Data Reconstruction Attack Client Gradient Edited Model

May 27, 2022

DELTA: Diverse Client Sampling for Fasting Federated Learning
Lin Wang, YongXin Guo, Tao Lin, Xiaoying Tang
Delta XAI Client Sampling Unbiased Sampling Client Gradient

January 28, 2022

Gradient Masked Averaging for Federated Learning
Irene Tenison, Sai Aravind Sreeramadas, Vaikkunth Mugunthan, Edouard Oyallon, Irina Rish, Eugene Belilovsky
Natural Gradient Distribution Generalization Gradient Update Skewed Non IID Client Gradient

January 10, 2022

Communication-Efficient Federated Learning with Accelerated Client Gradient
Geeho Kim, Jinkyu Kim, Bohyung Han
Convergence Rate Global Model Learning Framework Client Gradient Unstable Convergence

December 14, 2021

Federated Learning for Face Recognition with Gradient Correction
Yifan Niu, Weihong Deng
Face Recognition Softmax Function Unconstrained Face Recognition Client Gradient