Mitigating the effects of non-IID data in federated learning with a self-adversarial balancing method.
Anastasiya DanilenkaPublished in: FedCSIS (2023)
Keyphrases
- synthetic data
- prior knowledge
- data sets
- input data
- test data
- noisy data
- learning algorithm
- learning scheme
- bayesian methods
- training samples
- database
- online learning
- unsupervised learning
- dynamic programming
- active learning
- learning process
- missing values
- preprocessing
- detection method
- spectral clustering
- data collection
- reinforcement learning
- data structure
- high accuracy
- feature space
- learned models
- distributed data
- machine learning
- training data
- information loss
- learning tasks
- statistical methods
- pairwise
- background knowledge
- segmentation method
- missing data
- data sources
- high dimensional data
- knowledge discovery
- image data