FedOVA: One-vs-All Training Method for Federated Learning with Non-IID Data.
Yuanshao ZhuChristos MarkosRuihui ZhaoYefeng ZhengJames J. Q. YuPublished in: IJCNN (2021)
Keyphrases
- synthetic data
- prior knowledge
- training process
- prior information
- test data
- training samples
- input data
- supervised learning
- data sets
- training phase
- data analysis
- unsupervised learning
- minimum error rate
- bayesian methods
- statistical methods
- data sources
- active learning
- learning algorithm
- preprocessing
- data points
- learning speed
- training examples
- missing data
- background knowledge
- noisy data
- missing values
- neural network
- learning stage
- distributed data
- similarity measure
- objective function
- digital libraries
- support vector machine
- pairwise
- clustering method
- probabilistic model
- feature set