FedProb: An Aggregation Method Based on Feature Probability Distribution for Federated Learning on Non-IID Data.
Do-Van NguyenAnh-Khoa TranKoji ZettsuPublished in: Big Data (2022)
Keyphrases
- probability distribution
- synthetic data
- prior knowledge
- test data
- missing data
- data sets
- noisy data
- data sources
- detection method
- input data
- significant improvement
- statistical methods
- objective function
- background knowledge
- similarity measure
- prior information
- active learning
- probabilistic model
- pairwise
- missing values
- training data
- clustering method
- bayesian methods
- learned models
- support vector machine
- learning scheme
- data collection
- feature set
- image data
- knowledge discovery
- preprocessing
- data analysis
- unsupervised learning
- learning process
- information loss
- learning algorithm
- neural network