TY - GEN
T1 - Differentially private convolutional neural networks with adaptive gradient descent
AU - Huang, Xixi
AU - Guan, Jian
AU - Zhang, Bin
AU - Qi, Shuhan
AU - Wang, Xuan
AU - Liao, Qing
N1 - Publisher Copyright:
© 2019 IEEE.
PY - 2019/6
Y1 - 2019/6
N2 - Deep learning achieves remarkable success in the fields of target detection, computer vision, natural language processing, and speech recognition. However, traditional deep learning models may suffer the privacy risk due to some training data involve sensitive information, such as the medical histories, location information and face images. Attackers can exploit the implicit information to recover the sensitive information from the training data. In order to protecting privacy of deep learning model, we develop a novel optimization algorithm called DPAGD-CNN for convolution neural network which cooperates differential privacy technique. Specifically, DPAGD-CNN allocates privacy budgets more carefully in each iteration, rather than assigning a fixed privacy budget per iteration. We theoretically prove that our approach can protect the privacy of training data and it achieves higher classification accuracy under the moderate privacy budget in the MNIST and CIFAR-10 datasets.
AB - Deep learning achieves remarkable success in the fields of target detection, computer vision, natural language processing, and speech recognition. However, traditional deep learning models may suffer the privacy risk due to some training data involve sensitive information, such as the medical histories, location information and face images. Attackers can exploit the implicit information to recover the sensitive information from the training data. In order to protecting privacy of deep learning model, we develop a novel optimization algorithm called DPAGD-CNN for convolution neural network which cooperates differential privacy technique. Specifically, DPAGD-CNN allocates privacy budgets more carefully in each iteration, rather than assigning a fixed privacy budget per iteration. We theoretically prove that our approach can protect the privacy of training data and it achieves higher classification accuracy under the moderate privacy budget in the MNIST and CIFAR-10 datasets.
KW - Convolutional neural network
KW - Differential privacy
KW - Gradient descent
KW - Privacy-preserving
UR - https://www.scopus.com/pages/publications/85114777978
U2 - 10.1109/DSC.2019.00105
DO - 10.1109/DSC.2019.00105
M3 - 会议稿件
AN - SCOPUS:85114777978
T3 - Proceedings - 2019 IEEE 4th International Conference on Data Science in Cyberspace, DSC 2019
SP - 642
EP - 648
BT - Proceedings - 2019 IEEE 4th International Conference on Data Science in Cyberspace, DSC 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 4th IEEE International Conference on Data Science in Cyberspace, DSC 2019
Y2 - 23 June 2019 through 25 June 2019
ER -