DFRD: Data-Free Robustness Distillation for Heterogeneous Federated Learning.
Kangyang LuoShuai WangYexuan FuXiang LiYunshi LanMing GaoPublished in: NeurIPS (2023)
Keyphrases
- data sets
- prior knowledge
- data processing
- database
- data sources
- learning process
- data collection
- original data
- data structure
- digital libraries
- data analysis
- heterogeneous databases
- raw data
- sensor data
- learning algorithm
- missing data
- learning tasks
- spatial data
- noisy data
- statistical analysis
- unsupervised learning
- supervised learning
- bayesian methods
- distributed data
- heterogeneous data
- complex data
- data quality
- hidden variables
- semantically rich
- data distribution
- experimental data
- attribute values
- learning systems
- synthetic data
- input data
- knowledge discovery
- data points
- probabilistic model
- training data
- decision trees