machine learning
联邦近端
Federated ProximalA variant of Federated Learning optimization that adds a proximal term to the local objective function to constrain the local model updates, limiting the impact of non-IID data and client heterogeneity. / 一种联邦学习优化变体，在本地目标函数中添加近端项以约束本地模型更新，从而限制非独立同分布数据和客户端异质性的影响。
- Knowledge type
- 工艺技术
- First proposed or discovered
- 2017
- Proposer
- Li, Qiang, et al.
- Creator
- 农场主
- Last updated
- 2026-11-03 16:05:45
Sub-knowledge points
Local model drift limitationHeterogeneous client optimization stability
Higher-order knowledge points
Federated Learning / 实际应用Optimization Theory
Main parameters and performance indicators
mu: Proximal parameter controlling the strength of the regularization termlambda: Weighting parameter for the proximal term in the objective function
Open problems
No unresolved problem recorded yet
Related literature
ranked by community starsTrustworthy Federated Clinical Intelligence with Homomorphic Encryption and Explainable AIOpen source
Related products and links
No item recorded yet
Creation and modification trail
1 recordsRecord title: Federated Proximal.
Node update feedback
review required