Concealing Backdoor Model Updates in Federated Learning by Trigger-Optimized Data Poisoning.
Yujie ZhangNeil GongMichael K. ReiterPublished in: CoRR (2024)
Keyphrases
- learning models
- prior knowledge
- experimental data
- data sets
- data sources
- empirical data
- data collection
- test data
- learning scheme
- learning algorithm
- learning phase
- learned models
- data structure
- probability distribution
- image data
- simulation data
- expert knowledge
- learning tasks
- sensory data
- learning mechanism
- mathematical model
- statistical methods
- training data
- probabilistic model
- data points
- bayesian methods
- accurate models
- data analysis
- hidden markov models
- input data
- raw data
- hidden variables
- database
- distributed data
- semi supervised
- xml documents
- measured data
- databases
- high dimensional data