Privacy-Preserving Deep Learning via Additively Homomorphic Encryption

Phong, Le Trieu; Aono, Yoshinori; Hayashi, Takuya; Wang, Lihua; Moriai, Shiho

doi:10.1109/tifs.2017.2787987

Cited by 975 publications

(398 citation statements)

References 17 publications

Supporting

Mentioning

322

Contrasting

Unclassified

Order By: Relevance

“…One approach with HE has been proposed for privacy-preserving weight transmission for multiple owners who wish to apply a machine learning method over combined data sets [21,22]. However, this approach can not be applied to network training in the encrypted domain.…”

Section: Privacy-preserving Machine Learningmentioning

confidence: 99%

Privacy-Preserving Deep Neural Networks with Pixel-Based Image Encryption Considering Data Augmentation in the Encrypted Domain

Sirichotedumrong

Maekawa

Kinoshita

et al. 2019

2019 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

We present a novel privacy-preserving scheme for deep neural networks (DNNs) that enables us not to only apply images without visual information to DNNs for both training and testing but to also consider data augmentation in the encrypted domain for the first time. In this paper, a novel pixel-based image encryption method is first proposed for privacy-preserving DNNs. In addition, a novel adaptation network is considered that reduces the influence of image encryption. In an experiment, the proposed method is applied to a well-known network, ResNet-18, for image classification. The experimental results demonstrate that conventional privacy-preserving machine learning methods including the state-of-the-arts cannot be applied to data augmentation in the encrypted domain and that the proposed method outperforms them in terms of classification accuracy.

show abstract

Section: Privacy-preserving Machine Learningmentioning

confidence: 99%

Privacy-Preserving Deep Neural Networks with Pixel-Based Image Encryption Considering Data Augmentation in the Encrypted Domain

Sirichotedumrong

Maekawa

Kinoshita

et al. 2019

2019 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

show abstract

“…However, although the data is not explicitly shared in the original format, it is still possible for adversaries to reconstruct the raw data approximately, especially when the architecture and parameters are not completely protected. In addition, FL can expose intermediate results such as parameter updates from an optimization algorithm like stochastic gradient descent (SGD), and the transmission of these gradients may actually leak private information [9] when exposed together with a data structure such as image pixels. In addition, the existence of malevolent users may induce further security issues.…”

Section: Introductionmentioning

confidence: 99%

On Safeguarding Privacy and Security in the Framework of Federated Learning

Ding

et al. 2020

IEEE Network

250

View full text Add to dashboard Cite

Motivated by the advancing computational capacity of wireless end-user equipment (UE), as well as the increasing concerns about sharing private data, a new machine learning (ML) paradigm has emerged, namely federated learning (FL). Specifically, FL allows a decoupling of data provision at UEs and ML model aggregation at a central unit. By training model locally, FL is capable of avoiding data leakage from the UEs, thereby preserving privacy and security to some extend. However, even if raw data are not disclosed from UEs, individual's private information can still be extracted by some recently discovered attacks in the FL architecture. In this work, we analyze the privacy and security issues in FL, and raise several challenges on preserving privacy and security when designing FL systems. In addition, we provide extensive simulation results to illustrate the discussed issues and possible solutions.

show abstract

“…Adversarial training [94] Very easy to understand and implement It depends upon the sample size in the training phase Scalable and have the ability to handle the complex dataset Defense distillation [80] Sample and have the defense ability Difficult to converge and high complexity Ensemble method [95] Model-independent, good generalization Do not rebut the training data and computation overhead Differential Privacy [96] Preserves the privacy of training and learning data It also affects legitimate data and model-independent Low overhead, low complexity Homomorphic Encryption [97] Maintains security and privacy of data and simple It increases the data size and extensive computation overhead as well as to detect information security threats. Restricted Boltzmann Machines (RBMs) are also used for the same purpose, but we cannot find much study using this technique for security purposes.…”

Section: Countermeasure Methods Advantages Disadvantagesmentioning

confidence: 99%

A Review of Deep Learning Security and Privacy Defensive Techniques

Tariq

Memon

Ahmed

et al. 2020

Mobile Information Systems

View full text Add to dashboard Cite

In recent past years, Deep Learning presented an excellent performance in different areas like image recognition, pattern matching, and even in cybersecurity. The Deep Learning has numerous advantages including fast solving complex problems, huge automation, maximum application of unstructured data, ability to give high quality of results, reduction of high costs, no need for data labeling, and identification of complex interactions, but it also has limitations like opaqueness, computationally intensive, need for abundant data, and more complex algorithms. In our daily life, we used many applications that use Deep Learning models to make decisions based on predictions, and if Deep Learning models became the cause of misprediction due to internal/external malicious effects, it may create difficulties in our real life. Furthermore, the Deep Learning training models often have sensitive information of the users and those models should not be vulnerable and expose security and privacy. The algorithms of Deep Learning and machine learning are still vulnerable to different types of security threats and risks. Therefore, it is necessary to call the attention of the industry in respect of security threats and related countermeasures techniques for Deep Learning, which motivated the authors to perform a comprehensive survey of Deep Learning security and privacy security challenges and countermeasures in this paper. We also discussed the open challenges and current issues.

show abstract

Privacy-Preserving Deep Learning via Additively Homomorphic Encryption

Cited by 975 publications

References 17 publications

Privacy-Preserving Deep Neural Networks with Pixel-Based Image Encryption Considering Data Augmentation in the Encrypted Domain

Privacy-Preserving Deep Neural Networks with Pixel-Based Image Encryption Considering Data Augmentation in the Encrypted Domain

On Safeguarding Privacy and Security in the Framework of Federated Learning

A Review of Deep Learning Security and Privacy Defensive Techniques

Contact Info

Product

Resources

About