On the Utility Recovery Incapability of Neural Net-based Differential Private Tabular Training Data Synthesizer under Privacy Deregulation.
Yucong LiuChi-Hua WangGuang ChengPublished in: CoRR (2022)
Keyphrases
- neural nets
- training data
- privacy preserving
- differential privacy
- private data
- privacy requirements
- private information
- privacy guarantees
- feed forward
- privacy preserving data mining
- neural network
- privacy preservation
- back propagation
- artificial neural networks
- homomorphic encryption
- learning algorithm
- training set
- personal information
- sensitive information
- learning tasks
- backpropagation neural networks
- supervised learning
- prior knowledge
- training process
- privacy policies
- data sharing
- decision trees
- data privacy
- data sets
- privacy protection
- counter propagation
- privacy concerns
- training examples
- hidden layer
- support vector machine
- control signals
- genetic algorithm
- class labels
- semi supervised learning
- decision making
- disclosure risk
- third party
- labeled data