scholarly journals Private Model Compression via Knowledge Distillation

2019 ◽  
Vol 33 ◽  
pp. 1190-1197 ◽  
Author(s):  
Ji Wang ◽  
Weidong Bao ◽  
Lichao Sun ◽  
Xiaomin Zhu ◽  
Bokai Cao ◽  
...  
Keyword(s):  
Mobile Devices ◽  
Differential Privacy ◽  
Service Providers ◽  
Sample Selection ◽  
Sensitive Information ◽  
Model Compression ◽  
Complex Models ◽  
Speed Up ◽  
Knowledge Distillation ◽  
And Control

The soaring demand for intelligent mobile applications calls for deploying powerful deep neural networks (DNNs) on mobile devices. However, the outstanding performance of DNNs notoriously relies on increasingly complex models, which in turn is associated with an increase in computational expense far surpassing mobile devices’ capacity. What is worse, app service providers need to collect and utilize a large volume of users’ data, which contain sensitive information, to build the sophisticated DNN models. Directly deploying these models on public mobile devices presents prohibitive privacy risk. To benefit from the on-device deep learning without the capacity and privacy concerns, we design a private model compression framework RONA. Following the knowledge distillation paradigm, we jointly use hint learning, distillation learning, and self learning to train a compact and fast neural network. The knowledge distilled from the cumbersome model is adaptively bounded and carefully perturbed to enforce differential privacy. We further propose an elegant query sample selection method to reduce the number of queries and control the privacy loss. A series of empirical evaluations as well as the implementation on an Android mobile device show that RONA can not only compress cumbersome models efficiently but also provide a strong privacy guarantee. For example, on SVHN, when a meaningful (9.83,10−6)-differential privacy is guaranteed, the compact model trained by RONA can obtain 20× compression ratio and 19× speed-up with merely 0.97% accuracy loss.

scholarly journals Simplified ciphertext-policy attribute-based encryption scheme with attribute level collusion resistance for cloud storage

2021 ◽  
Vol 2089 (1) ◽  
pp. 012010
Author(s):  
Naresh Vurukonda ◽  
Venkateshwarlu Velde ◽  
M. TrinathBasu ◽  
P. Tejasri
Keyword(s):  
Service Providers ◽  
Sensitive Information ◽  
Privacy Policy ◽  
Collusion Resistance ◽  
Cloud Data ◽  
Attribute Based Encryption ◽  
Cloud Server ◽  
Ciphertext Policy ◽  
And Control ◽  
Cloud Applications

Abstract The core objective of our paper is that to overcome the challenges of secure data sharing problem in cloud where users can access data if they acquire a certain set of attributes or credentials. At present, there is only one way for implementing such procedures is by employing a reliable server to mediate access control, store the data. The service providers have whole control over cloud applications, client’s data and hardware. However, if the data stored in the cloud server is conceded, then the data confidentiality will be compromised. With simplified CP-ABE data is confidential even though storage server is un trusted. Hence privacy of cloud data is enhanced by proposing a methodology that uses a privacy policy, encrypts the data based upon various group of attributes related to the user. The decryption of the data can be done only when his/her attributes match with the privacy policy. Simplified CP-ABE makes it easier to secure, broadcast and control the access of sensitive information, specifically across the cloud’s server and decrease collision resistance.

scholarly journals Blockchain-Based Privacy-Preserving System for Genomic Data Management Using Local Differential Privacy

Electronics ◽  
2021 ◽  
Vol 10 (23) ◽  
pp. 3019
Author(s):  
Young-Hoon Park ◽  
Yejin Kim ◽  
Junho Shim
Keyword(s):  
Data Management ◽  
Differential Privacy ◽  
Genomic Data ◽  
Third Party ◽  
Data Management System ◽  
Sensitive Information ◽  
Healthcare Data ◽  
Large Size ◽  
Gene Data ◽  
And Control

The advances made in genome technology have resulted in significant amounts of genomic data being generated at an increasing speed. As genomic data contain various privacy-sensitive information, security schemes that protect confidentiality and control access are essential. Many security techniques have been proposed to safeguard healthcare data. However, these techniques are inadequate for genomic data management because of their large size. Additionally, privacy problems due to the sharing of gene data are yet to be addressed. In this study, we propose a secure genomic data management system using blockchain and local differential privacy (LDP). The proposed system employs two types of storage: private storage for internal staff and semi-private storage for external users. In private storage, because encrypted gene data are stored, only internal employees can access the data. Meanwhile, in semi-private storage, gene data are irreversibly modified by LDP. Through LDP, different noises are added to each section of the genomic data. Therefore, even though the third party uses or exposes the shared data, the owner’s privacy is guaranteed. Furthermore, the access control for each storage is ensured by the blockchain, and the gene owner can trace the usage and sharing status using a decentralized application in a mobile device.

scholarly journals Effective Privacy-Preserving Collection of Health Data from a User’s Wearable Device

2020 ◽  
Vol 10 (18) ◽  
pp. 6396
Author(s):  
Jong Wook Kim ◽  
Su-Mee Moon ◽  
Sang-ug Kang ◽  
Beakcheol Jang
Keyword(s):  
Differential Privacy ◽  
Service Providers ◽  
Healthcare Services ◽  
Wearable Devices ◽  
Privacy Preserving ◽  
Health Data ◽  
Experimental Results ◽  
Sensitive Information ◽  
Privacy Concerns ◽  
Primary Means

The popularity of wearable devices equipped with a variety of sensors that can measure users’ health status and monitor their lifestyle has been increasing. In fact, healthcare service providers have been utilizing these devices as a primary means to collect considerable health data from users. Although the health data collected via wearable devices are useful for providing healthcare services, the indiscriminate collection of an individual’s health data raises serious privacy concerns. This is because the health data measured and monitored by wearable devices contain sensitive information related to the wearer’s personal health and lifestyle. Therefore, we propose a method to aggregate health data obtained from users’ wearable devices in a privacy-preserving manner. The proposed method leverages local differential privacy, which is a de facto standard for privacy-preserving data processing and aggregation, to collect sensitive health data. In particular, to mitigate the error incurred by the perturbation mechanism of location differential privacy, the proposed scheme first samples a small number of salient data that best represents the original health data, after which the scheme collects the sampled salient data instead of the entire set of health data. Our experimental results show that the proposed sampling-based collection scheme achieves significant improvement in the estimated accuracy when compared with straightforward solutions. Furthermore, the experimental results verify that an effective tradeoff between the level of privacy protection and the accuracy of aggregate statistics can be achieved with the proposed approach.

scholarly journals Combine-Net: An Improved Filter Pruning Algorithm

Information ◽  
2021 ◽  
Vol 12 (7) ◽  
pp. 264
Author(s):  
Jinghan Wang ◽  
Guangyue Li ◽  
Wenzhao Zhang
Keyword(s):  
Structured Model ◽  
Compression Algorithms ◽  
Pruning Algorithm ◽  
Model Compression ◽  
The Neural Network ◽  
Empirical Determination ◽  
Knowledge Distillation ◽  
Resource Constrained Devices ◽  
Constrained Devices

The powerful performance of deep learning is evident to all. With the deepening of research, neural networks have become more complex and not easily generalized to resource-constrained devices. The emergence of a series of model compression algorithms makes artificial intelligence on edge possible. Among them, structured model pruning is widely utilized because of its versatility. Structured pruning prunes the neural network itself and discards some relatively unimportant structures to compress the model’s size. However, in the previous pruning work, problems such as evaluation errors of networks, empirical determination of pruning rate, and low retraining efficiency remain. Therefore, we propose an accurate, objective, and efficient pruning algorithm—Combine-Net, introducing Adaptive BN to eliminate evaluation errors, the Kneedle algorithm to determine the pruning rate objectively, and knowledge distillation to improve the efficiency of retraining. Results show that, without precision loss, Combine-Net achieves 95% parameter compression and 83% computation compression on VGG16 on CIFAR10, 71% of parameter compression and 41% computation compression on ResNet50 on CIFAR100. Experiments on different datasets and models have proved that Combine-Net can efficiently compress the neural network’s parameters and computation.

FISCAL EQUIVALENCE: PRINCIPLE AND PREDATION IN THE PUBLIC ADMINISTRATION OF JUSTICE

2021 ◽  
Vol 38 (1) ◽  
pp. 244-265
Author(s):  
Emily C. Skarbek
Keyword(s):  
Public Administration ◽  
Equivalence Principle ◽  
Justice System ◽  
Service Providers ◽  
The Political ◽  
The Public ◽  
Local Police ◽  
Administration Of Justice ◽  
Internal Pressures ◽  
And Control

AbstractFiscal equivalence in the public administration of justice requires local police and courts to be financed exclusively by the populations that benefit from their services. Within a polycentric framework, broad based taxation to achieve fiscal equivalence is a desirable principle of public finance because it conceptually allows for the provision of justice to be determined by constituent’s preferences, and increases the political accountability of service providers to constituents. However, the overproduction of justice services can readily occur when the benefits of the justice system are not enjoyed equally. Paradoxically, the same properties that make fiscal equivalence desirable by imposing restraint and control between constituents and local government also create internal pressures for agents of the state to engage in predatory, revenue-generating behavior.

Data-Free Ensemble Knowledge Distillation for Privacy-conscious Multimedia Model Compression

2021 ◽  
Author(s):  
Zhiwei Hao ◽  
Yong Luo ◽  
Han Hu ◽  
Jianping An ◽  
Yonggang Wen
Keyword(s):  
Multimedia Model ◽  
Model Compression ◽  
Knowledge Distillation

Revisiting knowledge distillation for light-weight visual object detection

2021 ◽  
Vol 43 (13) ◽  
pp. 2888-2898
Author(s):  
Tianze Gao ◽  
Yunfeng Gao ◽  
Yu Li ◽  
Peiyuan Qin
Keyword(s):  
Object Detection ◽  
Essential Element ◽  
Detection Algorithm ◽  
Positive Sample ◽  
Detection Methods ◽  
Visual Object ◽  
Light Weight ◽  
Model Compression ◽  
Novel Approach ◽  
Knowledge Distillation

An essential element for intelligent perception in mechatronic and robotic systems (M&RS) is the visual object detection algorithm. With the ever-increasing advance of artificial neural networks (ANN), researchers have proposed numerous ANN-based visual object detection methods that have proven to be effective. However, networks with cumbersome structures do not befit the real-time scenarios in M&RS, necessitating the techniques of model compression. In the paper, a novel approach to training light-weight visual object detection networks is developed by revisiting knowledge distillation. Traditional knowledge distillation methods are oriented towards image classification is not compatible with object detection. Therefore, a variant of knowledge distillation is developed and adapted to a state-of-the-art keypoint-based visual detection method. Two strategies named as positive sample retaining and early distribution softening are employed to yield a natural adaption. The mutual consistency between teacher model and student model is further promoted through a hint-based distillation. By extensive controlled experiments, the proposed method is testified to be effective in enhancing the light-weight network’s performance by a large margin.

scholarly journals Device-Based Security to Improve User Privacy in the Internet of Things †

Sensors ◽  
2018 ◽  
Vol 18 (8) ◽  
pp. 2664 ◽  
Author(s):  
Luis Belem Pacheco ◽  
Eduardo Pelinson Alchieri ◽  
Priscila Mendez Barreto
Keyword(s):  
Cloud Computing ◽  
Internet Of Things ◽  
Single Point ◽  
Data Access ◽  
Cloud Services ◽  
Sensitive Information ◽  
User Privacy ◽  
Privacy And Security ◽  
Iot Devices ◽  
And Control

The use of Internet of Things (IoT) is rapidly growing and a huge amount of data is being generated by IoT devices. Cloud computing is a natural candidate to handle this data since it has enough power and capacity to process, store and control data access. Moreover, this approach brings several benefits to the IoT, such as the aggregation of all IoT data in a common place and the use of cloud services to consume this data and provide useful applications. However, enforcing user privacy when sending sensitive information to the cloud is a challenge. This work presents and evaluates an architecture to provide privacy in the integration of IoT and cloud computing. The proposed architecture, called PROTeCt—Privacy aRquitecture for integratiOn of internet of Things and Cloud computing, improves user privacy by implementing privacy enforcement at the IoT devices instead of at the gateway, as is usually done. Consequently, the proposed approach improves both system security and fault tolerance, since it removes the single point of failure (gateway). The proposed architecture is evaluated through an analytical analysis and simulations with severely constrained devices, where delay and energy consumption are evaluated and compared to other architectures. The obtained results show the practical feasibility of the proposed solutions and demonstrate that the overheads introduced in the IoT devices are worthwhile considering the increased level of privacy and security.

Sign in / Sign up

Export Citation Format

Share Document