Preserving differential privacy in convolutional deep belief networks

Keyphrases

• Approximation by Polynomials 50%
• Biomedical Images 25%
• Chebyshev Expansion 25%
• Clinical Records 25%
• Convolutional Deep Belief Network 100%
• Deep Learning 50%
• Deep Neural Network 25%
• Differential Privacy 100%
• Energy-based 25%
• Error Bound 25%
• Functional Mechanism 25%
• Handwritten Digit Recognition 25%
• Health Social Network 25%
• Healthcare Domain 25%
• Highly Effective 25%
• Highly Sensitive 25%
• Human Activity Classification 25%
• Human Behavior Prediction 25%
• Main Idea 25%
• MNIST 25%
• Polynomial Representation 50%
• Preserving Privacy 25%
• Privacy Concerns 25%
• Recognition Task 25%
• Sensitivity Bounds 25%
• User Profile 25%

Computer Science

• Deep Learning 50%
• Deep Neural Network 25%
• Differential Privacy 100%
• Objective Function 50%
• Privacy Preserving 25%
• Representation of Polynomial 25%
• Scientific Study 25%
• Social Network 25%