DWFed: A statistical- heterogeneity-based dynamic weighted model aggregation algorithm for federated learning.
Aiguo ChenYang FuLingfu WangGuiduo DuanPublished in: Frontiers Neurorobotics (2022)
Keyphrases
- learning algorithm
- probabilistic model
- mathematical model
- objective function
- learning phase
- cost function
- goodness of fit
- optimization model
- input data
- algorithm employs
- np hard
- prior knowledge
- learned models
- estimation algorithm
- selection algorithm
- theoretical analysis
- recognition algorithm
- energy function
- learning mechanism
- automatically learned
- learning scheme
- classification algorithm
- computational complexity
- learning process
- em algorithm
- optimization algorithm
- similarity measure
- optimal solution
- fully connected
- k means
- dynamic programming
- expectation maximization
- segmentation algorithm
- bayesian framework
- prior information
- kalman filter
- detection algorithm
- learning models
- worst case
- parameter estimation
- multiple kernel
- hypothesis testing
- active learning
- voting scheme
- statistical methods