A two-phase half-async method for heterogeneity-aware federated learning.
Tianyi MaBingcheng MaoMing ChenPublished in: Neurocomputing (2022)
Keyphrases
- cost function
- synthetic data
- learning algorithm
- experimental evaluation
- preprocessing
- learning systems
- unsupervised learning
- high precision
- classification method
- mutual information
- training process
- support vector machine
- significant improvement
- prior knowledge
- similarity measure
- neural nets
- data sets
- optimization method
- learning scheme
- segmentation algorithm
- high accuracy
- supervised learning
- classification accuracy