Cloud-edge collaboration based computer vision inference mechanism by Boheng TANG, Xingang CHAI

“…The popularity of deep learning and cloud computing has promoted the widespread application of computer vision in various industries.However, centralized cloud inference services have problems such as high bandwidth resource consumption, image data privacy leakage, and high latency.It is hard that satisfy demand which requires diversified computer vision application.The dual gigabit upgrade of the communication network will promote depth collaboration of computer vision cloud-edge algorithms.Aiming to study the computer vision inference mechanism based on cloud-edge collaboration.Firstly, the advantages and disadvantages of the mainstream cloud and edge computer vision inference models in recent years were analyzed and explained, and on this basis, research on the cloud-edge collaborative computer vision inference model framework and deployment mechanism was carried out, model distributed reasoning model segmentation strategy, cloud-side collaborative network deployment optimization strategy was discussed in detail.In the end, the challenge and prospect of deep learning cloud-edge collaboration inference in future was discussed through data collaboration, network partition collaboration, and business function collaboration .…”
Get full text

Resource scheduling optimization in cloud-edge collaboration by Shuling WANG, Jie SUN, Peng WANG, Aidong YANG

“…With the enrichment and diversification of business types, low latency, high bandwidth, data privacy and high reliability have become common requirements.Edge computing, fog computing, distributed cloud, computing power network and other solutions have been proposed, and have triggered in-depth research and exploration in industry, academia and research.There is a consensus within and outside the industry on the view that “multi-level computing power distribution and collaboration of computing power will be the mainstream of computing power structure in the future”.The problems related to resource scheduling optimization, such as computing power management, allocation, scheduling, have also become the current research hotspot and key research direction.Therefore, for the future computing power supply structure, focuses on the latest progress of resource scheduling optimization in academia and industry, the current main methodology and engineering implementation architecture was summarized.And then, for the two typical cloud edge collaboration scenarios, the analysis was carried out from the perspective of scene splitting, scheduling objectives, and solutions in turn, and the resource scheduling optimization reference schemes that adapted to the characteristics of the scenarios were analyzed and discussed respectively.…”
Get full text

Privacy-protected crowd-sensed data trading algorithm by Yong ZHANG, Dandan LI, Lu HAN, Xiaohong HUANG

“…To solve the problem that data privacy leakage of participants under the crowd-sensed data trading model, a privacy-protected crowd-sensed data trading algorithm was proposed.Firstly, to achieve the privacy protection of participants, an aggregation scheme based on differential privacy was designed.Participants were no longer needed to upload raw data, but analyzed and calculated the collected data according to the task requirements, and then sent the analysis results to the platform after adding noise in accordance with the privacy budget allocated by the platform to protect their privacy.Secondly, in order to ensure the credibility of participants, a reputation model of participants was proposed.Finally, in order to encourage consumers and participants to participate in transactions, a data trading optimization model was constructed by considering the consumer’s constraint on the result deviation，the participant’s privacy leakage compensation and platform profit, and a POA based on genetic algorithm was proposed to solve the model.The simulation results show that the POA not only protects the privacy of participants, but also increases the profit of the platform by 29.27% and 20.45% compared to VENUS and DPDT, respectively.…”
Get full text

Discussion on Artificial Intelligence Safety and Ethical Issues by Chen Xinyu, Hui Tianfang, Li Yanlin, Yang Haoyuan

“…In addition, it is recommended to adopt measures such as data encryption and differential privacy to address data privacy and security issues. Regarding ethical considerations, this paper identifies the origins of algorithmic bias and argues for mitigating it through rigorous testing, validation, and regulatory frameworks. …”
Get full text

Efficient unpaired data validation and aggregation protocol in industrial Internet of things by MA Rong, FENG Tao

“…Within the context of an IIoT environment based on elliptic curve cryptography, homomorphic encryption was emploied to safeguard data privacy and a verification key management scheme was introduced, facilitating secure and efficient unpaired verification. …”
Get full text

Membership inference attack and defense method in federated learning based on GAN by Jiale ZHANG, Chengcheng ZHU, Xiaobing SUN, Bing CHEN

“…Aiming at the problem that the federated learning system was extremely vulnerable to membership inference attacks initiated by malicious parties in the prediction stage, and the existing defense methods were difficult to achieve a balance between privacy protection and model loss.Membership inference attacks and their defense methods were explored in the context of federated learning.Firstly, two membership inference attack methods called class-level attack and user-level attack based on generative adversarial network (GAN) were proposed, where the former was aimed at leaking the training data privacy of all participants, while the latter could specify a specific participant.In addition, a membership inference defense method in federated learning based on adversarial sample (DefMIA) was further proposed, which could effectively defend against membership inference attacks by designing adversarial sample noise addition methods for global model parameters while ensuring the accuracy of federated learning.The experimental results show that class-level and user-level membership inference attack can achieve over 90% attack accuracy in federated learning, while after using the DefMIA method, their attack accuracy is significantly reduced, approaching random guessing (50%).…”
Get full text

Federated Learning for Decentralized DDoS Attack Detection in IoT Networks by Yaser Alhasawi, Salem Alghamdi

“…Our approach prioritizes data privacy by processing data locally, thereby avoiding the need for central data collection, while enhancing detection efficiency. …”
Get full text

Safeguarding Mobile Users from Violation by Third-party Apps by Vusumuzi Malele, Kagiso Mphasane

“…Insecure third-party mobile applications (apps) can have a detrimental impact on mobile users in terms of information security and data privacy. Insufficient protection for third-party mobile apps platforms may result in harmful installations. …”
Get full text

Research on dynamic cooperative technology of manned and unmanned networked information system by Haijun YE, Guofeng WANG, Zhiyong FENG

“…To meet the needs of information security sharing and systematic dynamic cooperation of manned and unmanned formations and to build a dynamic communication network, firstly, a networked high-precision system time synchronization scheme was designed, using integrated network link management strategy and wireless channel on-demand preference algorithm, and the time synchronization accuracy could be up to 100 nanosecond level.Secondly, a data privacy protection and secure sharing solution for airborne networks was built, implementing communication data security of manned and unmanned formation, and analyzing and discussing the performance and security to prove that the proposed solution has a provable high security strength.Meanwhile, a dynamic cooperation model of manned and unmanned formation was designed to discover, authenticate and revoke each node in real time, perform legitimacy identity verification, optimize allocation scheme for multi-objective information acquisition through multi-machine collaborative construction, differentiated data was used to transmit synchronize data, save network resources, improve data transmission efficiency, and meet the demand of systematic cooperative operations.…”
Get full text

Secure data deduplication for Internet-of-things sensor networks based on threshold dynamic adjustment by Yuan Gao, Hequn Xian, Aimin Yu

“…The item response theory is adopted to determine the sensitivity of different data and their privacy score, which ensures the applicability of data privacy score. It can solve the problem that some users care little about the privacy issue. …”
Get full text

Application of federated learning in predicting breast cancer by Chai Jiarui

“…Federated learning provides a framework to protect data privacy, allowing multiple institutions to share model training without sharing the original data. …”
Get full text

Federated learning-based user access strategy and energy consumption optimization in cell-free massive MIMO network by Yuanyuan YAO, Yiqiu LIU, Sai HUANG, Chunyu PAN, Xuehua LI, Xin YUAN

“…To solve the problem that how users choose access points in cell-free massive multiple-input multiple-output (CF-mMIMO) network, a prioritized access strategy for poorer users based on channel coefficient ranking was proposed.First, users were evaluated and ranked for their channel quality and stability after channel sensing, and suitable access points were selected in sequence according to the order of the channel state information.Second, considering issues such as users' energy consumption and data security, a federal learning framework was used to enhance user's data privacy and security.Meanwhile, an alternating optimization variables algorithm based on energy consumption optimization was proposed to optimize the multi-dimensional variables, for the purpose of minimizing the total energy consumption of the system.Simulation results show that compared with the traditional user-centric in massive MIMO, the proposed access strategy can improve the average uplink reachable rate of users by 20%, and the uplink rate of users with poor channels can be double improved; in terms of energy consumption optimization, the total energy consumption can be reduced by much more than 50% after optimization.…”
Get full text

Research on the cooperative offloading strategy of sensory data based on delay and energy constraints by Peiyan YUAN, Saike SHAO, Ran WEI, Junna ZHANG, Xiaoyan ZHAO

“…The edge offloading of the internet of things (IoT) sensing data was investigated.Multiple edge servers cooperatively offload all or part of the sensing data initially sent to the cloud center, which protects data privacy and improves user experience.In the process of cooperative offloading, the transmission of the sensing data and the information exchange among edge servers will consume system resources, resulting in the cost of cooperation.How to maximize the offloading ratio of the sensing data while maintaining a low collaboration cost is a challenging problem.A joint optimization problem of sensing data offload ratio and cooperative scale satisfying the constraints of network delay and system energy consumption was formulated.Subsequently, a distributed alternating direction method of multipliers (ADMM) via constraint projection and variable splitting was proposed to solve the problem.Finally, simulation experiments were carried out on MATLAB.Numerical results show that the proposed method improved the network delay and energy consumption compared to the fairness cooperation algorithm (FCA), the distributed optimization algorithm (DOA), and multi-subtasks-to-multi-servers offloading scheme (MTMS) algorithm.…”
Get full text

Edge computing privacy protection method based on blockchain and federated learning by Chen FANG, Yuanbo GUO, Yifeng WANG, Yongjin HU, Jiali MA, Han ZHANG, Yangyang HU

“…Aiming at the needs of edge computing for data privacy, the correctness of calculation results and the auditability of data processing, a privacy protection method for edge computing based on blockchain and federated learning was proposed, which can realize collaborative training with multiple devices at the edge of the network without a trusted environment and special hardware facilities.The blockchain was used to endow the edge computing with features such as tamper-proof and resistance to single-point-of-failure attacks, and the gradient verification and incentive mechanism were incorporated into the consensus protocol to encourage more local devices to honestly contribute computing power and data to the federated learning.For the potential privacy leakage problems caused by sharing model parameters, an adaptive differential privacy mechanism was designed to protect parameter privacy while reducing the impact of noise on the model accuracy, and moments accountant was used to accurately track the privacy loss during the training process.Experimental results show that the proposed method can resist 30% of poisoning attacks, and can achieve privacy protection with high model accuracy, and is suitable for edge computing scenarios that require high level of security and accuracy.…”
Get full text

Methodological and Technological Advancements in E-Learning by Elias Dritsas, Maria Trigka

“…The survey addresses critical challenges such as the digital divide, data privacy, and resistance to adoption, offering evidence-based strategies to mitigate these issues. …”
Get full text

Exploring the Role of Innovative Technologies in Smart Cities by Kagontle Selwe, Costa Hofisi

“…The article identified the unique challenges posed by the very nature of smart city networks which include Cybersecurity Risks and Data Privacy Concerns. Despite the promising potential of smart city technologies to create a more efficient, sustainable, and citizen-focused urban environment, the challenges identified in this article should be addressed in order to unlock the full potential of innovative technologies for smart cities.…”
Get full text

Survey of Distributed Ledger Technology (DLT) for Secure and Scalable Computing by Shreya Girish Savadatti, Shruthi Krishnamoorthy, Radhakrishnan Delhibabu

“…It also delves into DLT’s regulatory landscape, addressing compliance, data privacy and governance, while assessing the potential for regulatory clarity and international cooperation. …”
Get full text

AI Horizons in Indian Healthcare: A Vision for Transformation and Equity by Neelesh Kapoor, S N Sanjana, Shubha B. Davalagi, P S Balu, Soumitra Sethia

“…While these advancements show promise, significant challenges persist, related to data privacy concerns and interoperability issues, including the need for robust ethical frameworks. …”
Get full text

A survey of security threats in federated learning by Yunhao Feng, Yanming Guo, Yinjian Hou, Yulun Wu, Mingrui Lao, Tianyuan Yu, Gang Liu

“…This points to the need for further research into defensive approaches to make federated learning a real solution for distributed machine learning paradigm with securing data privacy. Our survey provides a taxonomy of these threats and defense methods, describing the general situation of this vulnerability in federated learning. …”
Get full text

Efficient secure federated learning aggregation framework based on homomorphic encryption by Shengxing YU, Zhong CHEN

“…In order to solve the problems of data security and communication overhead in federated learning, an efficient and secure federated aggregation framework based on homomorphic encryption was proposed.In the process of federated learning, the privacy and security issues of user data need to be solved urgently.However, the computational cost and communication overhead caused by the encryption scheme would affect the training efficiency.Firstly, in the case of protecting data security and ensuring training efficiency, the Top-K gradient selection method was used to screen model gradients, reducing the number of gradients that need to be uploaded.A candidate quantization protocol suitable for multi-edge terminals and a secure candidate index merging algorithm were proposed to further reduce communication overhead and accelerate homomorphic encryption calculations.Secondly, since model parameters of each layer of neural networks had characteristics of the Gaussian distribution, the selected model gradients were clipped and quantized, and the gradient unsigned quantization protocol was adopted to speed up the homomorphic encryption calculation.Finally, the experimental results show that in the federated learning scenario, the proposed framework can protect data privacy, and has high accuracy and efficient performance.…”
Get full text