FedTSA: A Cluster-based Two-Stage Aggregation Method for Model-heterogeneous Federated Learning.
Boyu FanChenrui WuXiang SuPan HuiPublished in: CoRR (2024)
Keyphrases
- prior knowledge
- learning scheme
- learning mechanism
- mathematical model
- statistical model
- learning algorithm
- evaluation method
- neural nets
- sensitivity analysis
- probabilistic model
- objective function
- segmentation method
- modeling method
- theoretical analysis
- hybrid model
- linear model
- structural learning
- cost function
- learning process
- reinforcement learning
- high order
- prior information
- optimization method
- reconstruction method
- boosted decision trees
- bp neural network
- linear regression
- evaluation model
- unsupervised learning
- digital libraries
- gaussian distribution
- prediction model
- closed form
- test data
- energy function
- computational model
- detection method
- model free
- em algorithm
- automatically learned
- learned models
- statistical natural language processing
- learning models
- statistical methods
- parameter estimation
- generative model
- input data
- support vector machine
- dynamic programming
- significant improvement
- pairwise
- similarity measure
- image sequences