A Study of Ensemble Feature Selection and Adversarial Training for Malicious User Detection

The continuously booming of information technology has shed light on developing a variety of communication networks,multimedia,social networks and Internet of Things applications.However,users inevitably suffer from the intrusion of malicious users.Some studies focus on static characteristics of malicious users,which is easy to be bypassed by camouflaged malicious users.In this paper,we present a malicious user detection method based on ensemble feature selection and adversarial training.Firstly,the feature selection alleviates the dimension disaster problem and achieves more accurate classification performance.Secondly,we embed features into the multidimensional space and aggregate it into a feature map to encode the explicit content preference and implicit interaction preference.Thirdly,we use an effective ensemble learning which could avoid over-fitting and has good noise resistance.Finally,we propose a datadriven neural network detection model with the regularization technique adversarial training to deeply analyze the characteristics.It simplifies the parameters,obtaining more robust interaction features and pattern features.We demonstrate the effectiveness of our approach with numerical simulation results for malicious user detection,where the robustness issues are notable concerns.

Optimal Cooperative Spectrum Sensing Based on Butterfly Optimization Algorithm

Since the introduction of the Internet of Things(IoT),several researchers have been exploring its productivity to utilize and organize the spectrum assets.Cognitive radio(CR)technology is characterized as the best aspirant for wireless communications to augment IoT competencies.In the CR networks,secondary users(SUs)opportunistically get access to the primary users(PUs)spectrum through spectrum sensing.The multipath issues in the wireless channel can fluster the sensing ability of the individual SUs.Therefore,several cooperative SUs are engaged in cooperative spectrum sensing(CSS)to ensure reliable sensing results.In CSS,security is still a major concern for the researchers to safeguard the fusion center(FC)against abnormal sensing reports initiated by the malicious users(MUs).In this paper,butterfly optimization algorithm(BOA)-based soft decision method is proposed to find an optimized weighting coefficient vector correlated to the SUs sensing notifications.The coefficient vector is utilized in the soft decision rule at the FC before making any global decision.The effectiveness of the proposed scheme is compared for a variety of parameters with existing schemes through simulation results.The results confirmed the supremacy of the proposed BOA scheme in both the normal SUs’environment and when lower and higher SNRs information is carried by the different categories of MUs.

Reputation-Based Cooperative Spectrum Sensing Algorithm for Mobile Cognitive Radio Networks

Identifying malicious users accurately in cognitive radio networks(CRNs) is the guarantee for excellent detection performance. However, existing algorithms fail to take the mobility of secondary users into consideration. If applied directly in mobile CRNs, those conventional algorithms would overly punish reliable users at extremely bad or good locations, leading to an obvious decrease in detection performance. To overcome this problem, we divide the whole area of interest into several cells to consider the location diversity of the network. Each user's reputation score is updated after each sensing slot and is used for identifying whether it is malicious or not. If so, it would be removed away. And then our algorithm assigns users in cells with better channel conditions, i.e. larger signal-to-noise ratios(SNRs), with larger weighting coefficients, without requiring the prior information of SNR. Detailed analysis about the validity of our algorithm is presented. The simulation results show that in a CRN with 60 mobile secondary users, among which, 18 are malicious, our solution has an improvement of detection probability by 0.97-d B and 3.57-d B when false alarm probability is 0.1, compared with a conventional trust-value-based algorithm and a trusted collaborative spectrum sensing for mobile CRNs, respectively.

Access control scheme with tracing for outsourced databases

To manage dynamic access control and deter pi- rate attacks on outsourced databases, a dynamic access control scheme with tracing is proposed. In our scheme, we introduce the traitor tracing idea into outsource databases, and employ a polynomial function and filter function as the basic means of constructing encryption and decryption procedures to reduce computation, communication, and storage overheads. Compared to previous access control schemes for outsourced databases, our scheme can not only protect sensitive data from leaking and perform scalable encryption at the server side without shipping the outsourced data back to the data owner when group membership is changed, but also provide trace-and-revoke features. When malicious users clone and sell their decryption keys for profit, our scheme can trace the decryption keys to the malicious users and revoke them. Furthermore, our scheme avoids massive message exchanges for establishing the decryption key between the data owner and the user. Compared to previously proposed publickey traitor tracing schemes, our scheme can simultaneously achieve full collusion resistance, full recoverability, full revocation, and black-box traceability. The proof of security and analysis of performance show that our scheme is secure and efficient.