Secure Aggregation in Federated Learning is not Private: Leaking User Data at Large Scale through Model Modification.
Joshua C. ZhaoAtul SharmaAhmed Roushdy ElkordyYahya H. EzzeldinSalman AvestimehrSaurabh BagchiPublished in: CoRR (2023)
Keyphrases
- prior knowledge
- learning models
- probability distribution
- experimental data
- data sets
- learned models
- accurate models
- test data
- input data
- learning process
- database
- data sources
- data structure
- end users
- computational model
- background knowledge
- user input
- sensory data
- bayesian methods
- training data
- hidden variables
- statistical methods
- security model
- learning scheme
- user privacy
- user model
- network structure
- user preferences
- user interaction
- distributed systems
- knowledge discovery
- data points
- probabilistic model
- user interface
- bayesian networks