Search

Designing energy-efficient computing solution for the implementation of AI algorithms in edge devices remains a challenge. Yang et al. proposes a decentralized brain-inspired computing method enabling multiple edge devices to collaboratively train a global model without a fixed central coordinator.

The emergence of Internet of Things (IoT) and high throughput satellite communication networks enables the capability of anytime, anywhere environment monitoring and sensing. A key challenge of satellite-based IoT is to enhance spectrum and energy efficiency so as to meet the ever-increasing demand for satellite bandwidth and dynamic access of a massive number of IoT terminals. In this paper, we propose a novel power control algorithm for IoT terminals being deployed in satellite-based IoT systems where some terrestrial base station is available to acquire IoT devices' information as well as to perform resource management. We adopted the Poisson point process (PPP) theory to formulate the model for this power optimization problem. The PPP theory is applied to evaluate the distance distribution of random IoT devices in this satellite-based networks. Optimal power control scheme can be obtained by taking into consideration user distribution and signal interference plus noise ratio (SINR) demand for various IoT terminals. In addition, due to the complexity of the objective function of power control deduced by the PPP theory, we utilize the pattern search method to identify an optimal solution in global area. Furthermore, we provide numerical results from various perspectives including user rates and energy efficiency to testify the performances of our power proposal.

Next-generation mobile communication networks promise the support for vast number of IoT devices with strong demand for spectrum access. The cater for the continuous growth of IoT applications, one of the challenges to mobile communication network providers is to allow dynamic spectrum access to a gigantic number of densely distributed and low-power IoT networks. In this paper, a novel dynamic spectrum access method is proposed based on spectrum trading for distributed IoT devices. The notion of access benefit is introduced which is based on the purchased band number and modulation mode provided to the mobile user. In the proposed solution, IoT users aim achieve optimization of their spectrum access by ensuring that the spectrum price will not exceed the access benefits from the spectrum trading by solving the spectrum optimization function. Due to the complexity of optimization objective function, pattern search algorithm is utilized to complete the final spectrum allocation optimization. Numerical results are also provided to testify the performances of the proposed spectrum optimization method.

Energy Efficiency (EE) is a significant problem for Cognitive Radio (CR) network. Recently, more interest has focused on EE optimization problem in wireless-powered communication. In conventional CR network, spectrum sensing and limited battery capacity may decrease system performance. In this article, a Non-Orthogonal Multiple Access (NOMA) system with Simultaneous Wireless Information and Power Transfer (SWIPT) for CR network is studied. The frame structure is designed with two subslots. In the downlink subslot, the Secondary Users (SUs) harvest wireless energy from Radio Frequency (RF) signals and sense the spectrum state simultaneously. In the uplink subslot, SUs transmit their independent information to Base Station (BS). Two modes are considered in this article: overlay network and underlay network. A CR-NOMA system model is presented and the approximate expressions of EE for two modes are obtained. Based on the subslot allocation, two optimization problems aiming to maximize EE are formulated. In the overlay network, the constraints are transmit power and total transmission slot. In the underlay network, instead of sensing the spectrum, SUs utilize the channel with primary user (PU) simultaneously. Thus, the constraints of interference threshold and channel gain of PU are also taken into considered. The proposed optimization problems can be regarded as nonlinear fractional programming. The Dinkelbach method is used to transform the nonlinear fractional programming problems into the parametric ones. Simulation results show that there indeed exists a best downlink subslot to maximize the EE of CR-NOMA networks.

With the availability of high-throughput satellite services at affordable cost, terrestrial network providers make use of satellite links to extend their coverage to areas, where land-based communication infrastructures are prohibitively costly to implement. Due to the ever-increasing demand for bandwidth resulted from the rapid development of data-intensive services in recent years, one of the fundamental challenges for satellite communications is to continuously improve utilization efficiency of the scarce satellite spectrum. Cognitive satellite communications address the problem by providing mechanisms for terrestrial and satellite users to dynamically access idle bands of licensed satellite networks, hence enabling spectrum sharing between two satellite systems or between satellite and terrestrial systems. In this paper, we investigate a distributed technique of spectrum sharing for cognitive satellite networks. In order to cater for situations with incomplete decision information, the proposed scheme is based on the Bayesian equilibrium theory. We first develop a spectrum allocation scheme by studying the action strategy of terrestrial cognitive terminals in a distributed competition. A feasible spectrum allocation scheme that caters for cases of incomplete user information is then developed as an extension of the basic scheme by formulating the problem as a Cournot game model. By identifying the unique equilibrium in this model, optimal spectrum allocation for cognitive satellite networks can be achieved. Essential discussions and proofs for the rationality of this method and uniqueness of the equilibrium are provided. Numerical results are given to justify the claimed advantages.

With the rapid development of modern satellite communications, broadband satellite services are experiencing a period of remarkable growth in both user population and available bandwidth. Efficient spectrum management is required in order to meet the ever-increasing demand for broadband spectrum. In this paper, a spectrum allocation scheme for cognitive satellite networks is proposed to improve spectrum efficiency by addressing the situation that scarce spectrum resource is under-utilized while the overall demands of cognitive satellite users are not satisfied. The proposed scheme is devised to optimize bandwidth efficiency by achieving the Bayesian equilibrium through spectrum competition among cognitive satellite users. Due to the scarcity of spectrum resource, satellite systems have to eliminate some of the cognitive users' transmission information types in order to fulfill the spectrum needs of priority users. After declining one or several kinds of modulation modes, cognitive users can update their spectrum lists. With the increase in number of cognitive users and their corresponding transmission information types, the complexity of the optimization problem increases significantly and becomes a computation burden to cognitive satellite systems. Therefore, in this paper, we propose a simplified solution for spectrum resource optimization in a computation-efficient manner. Through rounds of eliminating operations, the satellite systems can identify the final Bayesian equilibrium as an optimal spectrum allocation strategy. We also provide proofs for the existence of Bayesian equilibrium. Numerical results are given to evaluate the performance of our proposed method from diverse perspectives.

The Quality of Service (QoS)-based personalized web service recommendations have been gaining increasing popularity due to its ability to assist users in finding high quality web services. For this purpose, Collaborative Filtering (CF)-based technique has been a useful approach in that it is able to predict with high accuracy the QoS values of web services which are not invoked by the users. The basic idea behind CF-based techniques is that they identify users with similar QoS experiences and predict their QoS requirements on web services accordingly. However, as the calculation of QoS values and user similarity require parameters which may contain privacy sensitive information, users may not trust the server that provides such third-party recommendations. In general, users are usually not willing to disclose such information to a third-party as it contains their tastes and preferences as well as experiences. Therefore the main challenge is to address the need for providing accurate web service recommendations to users while preserving their privacy from any third party server, as well as to protect the privacy of individual users from one another. To tackle this challenge, we propose a new protocol for privacy preserving web service recommendation where an untrusted recommendation server is able to provide the recommendation without disclosing any private information of individual users, and with negligible loss of accuracy of QoS values. We present both privacy and experimental analysis to verify that our proposed method is secure and efficient in terms of performance.

Successful cyber-attacks are caused by the exploitation of some vulnerabilities in the software and/or hardware that exist in systems deployed in premises or the cloud. Although hundreds of vulnerabilities are discovered every year, only a small fraction of them actually become exploited, thereby there exists a severe class imbalance between the number of exploited and non-exploited vulnerabilities. The open source national vulnerability database, the largest repository to index and maintain all known vulnerabilities, assigns a unique identifier to each vulnerability. Each registered vulnerability also gets a severity score based on the impact it might inflict upon if compromised. Recent research works showed that the cvss score is not the only factor to select a vulnerability for exploitation, and other attributes in the national vulnerability database can be effectively utilized as predictive feature to predict the most exploitable vulnerabilities. Since cybersecurity management is highly resource savvy, organizations such as cloud systems will benefit when the most likely exploitable vulnerabilities that exist in their system software or hardware can be predicted with as much accuracy and reliability as possible, to best utilize the available resources to fix those first. Various existing research works have developed vulnerability exploitation prediction models by addressing the existing class imbalance based on algorithmic and artificial data resampling techniques but still suffer greatly from the overfitting problem to the major class rendering them practically unreliable. In this research, we have designed a novel cost function feature to address the existing class imbalance. We also have utilized the available large text corpus in the extracted dataset to develop a custom-trained word vector that can better capture the context of the local text data for utilization as an embedded layer in neural networks. Our developed vulnerability exploitation prediction models powered by a novel cost function and custom-trained word vector have achieved very high overall performance metrics for accuracy, precision, recall, F1-Score and AUC score with values of 0.92, 0.89, 0.98, 0.94 and 0.97, respectively, thereby outperforming any existing models while successfully overcoming the existing overfitting problem for class imbalance.

There is a growing trend recently in big data analysis that focuses on behavior interiors, which concern the semantic meanings (e.g., sentiment, controversy, and other state-dependent factors) in explaining the human behaviors from psychology, sociology, cognitive science, and so on, rather than the data per se as in the case of exterior dimensions. It is more intuitive and much easier to understand human behaviors with less redundancy in concept by exploring the behavior interior dimensions, compared with directly using behavior exteriors. However, they usually approach from a unidimensional perspective with a lack of a sense of interrelatedness. Thus, integrating multiple behavior dimensions together into some numerical measures to form a more comprehensive view for subsequent prediction processes becomes a pivotal issue. Moreover, these studies usually focus on the magnitude but neglect the associated temporal features. In this paper, we propose a behavior interior dimension-based neighborhood collaborative filtering method for the top-N hashtag adoption frequency prediction that takes into account the interdependence in temporal dynamics. Our proposed approach couples the similarity in user preference and their impact propagation, by integrating the linear threshold model and the enhanced CF model based on behavior interiors. Experiments on Twitter demonstrate that the behavior-interior-aware CF models achieve better adoption prediction results than the state-of-the-art methods, and the joint consideration of similarity in user preference and their impact propagation results in a significant improvement than treating them separately.