Synthetic data shuffling accelerates the convergence of federated learning under data heterogeneity.
Bo LiYasin EsfandiariMikkel N. SchmidtTommy Sonne AlstrømSebastian U. StichPublished in: Trans. Mach. Learn. Res. (2024)
Keyphrases
- synthetic data
- data sets
- data analysis
- small number
- raw data
- database
- data sources
- active learning
- learning algorithm
- image data
- databases
- prior knowledge
- learning models
- convergence rate
- original data
- data mining techniques
- online learning
- data processing
- human experts
- spatial data
- knowledge acquisition
- mri data
- distributed data
- neural network
- labeled data
- data collection
- digital libraries
- real world
- knowledge discovery
- end users
- social networks
- learning process