With the rapid development of cloud computing, an increasing number of organizations and individuals store and share their data on clouds. While cloud storage brings people convenience, it also brings a series of problems at the same time. For example, the data uploaded to the cloud may be tampered and damaged. Therefore, shared data auditing has become an important issue in the field of cloud storage, attracting the attention of researchers. However, the state-of-the-art schemes cannot fully meet the performance and security requirements. Therefore, this paper proposes a public audit shared data protocol for cloud storage by using blockchain and Rank-based Merkle AVL tree (RB-MHT) to achieve privacy preserving and batch auditing to reduce system overhead as well as keeping the security of modification record in this blockchain-based scheme. For privacy preserving, the auditing signature is only related to group management during the audit process and data is blind by a random value. Furthermore, the security of the scheme is verified and its performance is evaluated through implementation. The results demonstrate that the proposed scheme is secure and efficient.

Under the background of extensive interconnection of power terminals, the traditional security and stability control architecture gradually shifts to distributed Special Protection System(SPS) with terminal intelligent collaboration. However, the identity authentication mechanism between the security and stability control terminals of the smart grid is not yet complete, thus it has become an irresistible trend to establish the mutual authentication and trust mechanism between the security and stability control terminals so as to avoid power system from cyber attacks as far as possible. This paper first analyzes the current status of SPS and the security risks SPS faced due to lack of targeted identity authentication mechanisms. On this basis, the alliance blockchain and the appropriate consensus mechanism are selected, and an identity authentication scheme based on blockchain technology that meets the characteristics of SPS is proposed. The distributed identity authentication scheme of blockchain technology uses hash chains and a distributed storage scheme to achieve lightweight authentication. Finally, the efficiency and security of the scheme are analyzed. The identity authentication mechanism proposed in this paper meets the high requirements of the power system and can defend against several common network attacks such as man-in-the-middle attacks and DoS attacks, thus ensuring the safe and stable operation of the power system.

At present, machine learning based Android malware detection approaches has the problem of model aging. Malware is constantly changing and evolving rapidly with time, which leads to concept drift. Concept drift makes underlying data distribution change over time, which violates the machine learning assumption that the data distribution is stable. In order to alleviate the problem of model aging, a confidence-guided hybrid malware detection system is proposed. By analyzing the credibility and confidence of the predicted results of heterogeneous models, this system can break through the problem that the heterogeneous models could not cooperate with each other. An open hybrid detection platform is established to mitigate concept drift. Experiments show that hybrid Android malware detection system is effective. In an evaluation with 66 000 applications, SVM model and random forest model have their own advantages and disadvantages. Hybrid Android malware detection system can improve the prediction effect on the basis of one single model.

To solve the data race problem in Android multi-threaded programs, a model of concurrent behavior detection based on Pi calculus is proposed in this paper. The extended Pi calculus is used to model Android life cycle and multi-threaded framework, and a formal behavior model is obtained. Moreover, the detection model is constructed by abstracting security constraints into formal IF-THEN rules and using the properties of Pi calculus to perform process calculus and migration. The dynamic detection and static detection are combined in the detection model by the same way, and concurrent behavior detection algorithm and data race detection method are given. The analysis and experiment results show that this method has linear time and space complexity. Compared with other methods, the detection accuracy is improved without sacrificing the detection efficiency.

The server is a fundamental facility for today’s information systems, cloud data storage and processing. CPU is the core element of the server. The current CPU circuit is extremely large in scale and complicated in production process, and its design, packaging and manufacturing are heavily dependent on foreign technologies and manufacturers. How to ensure the security and credibility of the processor chip is a key to network security and information security. But till today, credible research on the hardware behavior trustworthiness of CPU chip has not aroused sufficient attention. This paper first gives the concept of “trusted CPU chip”, combing the security risks faced by CPU chips in recent years. On this basis, the implementation principle of trusted CPU chip based on Tsinghua University DSC technology and its server trusted framework are proposed. Finally, this paper explores the significance of the trusted CPU chip and its server trusted framework in the current hardware security scenarios and the extended applications in covering high security requirements.

Based on the unique pictograph and the structure of Chinese character, a new way to form automatic summarization is proposed in the paper, which includes text vector technique directing at Chinese stroke and an automatic summarizing model. Stroke-based text vector codes the basic element of Chinese character and it highlights the specific characteristics of the word, which makes the relationship between words tightened. The corresponding text vector of Chinese word is gained by Skip-Gram model and optimized through Seq2Seq model. It solves the problem of long-sequence text information loss and the supplement of reversing information by using Bi-LSTM. Attention mechanism is used in encoder to weigh different effects of the input statement on decoder and meanwhile the use of Beam Search in the decoder optimizes the sequence of the results. The experiments based on LCSTS data set training model show the automatic summarization model can improve the quality and the readability of Chinese text summary.

In the task of Chinese event detection, there is a problem that domains are independent from each other, and data among domains can not be exchanged. It is necessary to label a large number of data for each domain. Based on previous studies, an open Chinese event detection method based on transfer learning is proposed in this paper. Two association hypotheses of trigger words are studied. The first one is that under the same event type, trigger words are strongly relevant in semantic space with each other. The other one is that among different event types, trigger words are also related with each other, but their relationship are weaker than those under the same event type. Based on the hypotheses, the relationship between candidate words and seed trigger words and the contextual features of candidate words are constructed with the help of external dictionaries. Then,the basic model and the transfer model of event detection are constructed by using convolutional neural network. Finally, only a small amount of tagged data is needed to detect events in the new domain. On ACE2005 Chinese event data set, this method only uses 20% of the data for trigger word recognition,and its effect can surpass the current mainstream method.

When dealing with imbalanced datasets, in order to reduce the impact of class overlapping on classification effect, and avoid over-fitting caused by over-sampling and information loss attributed to under-sampling, a Multi-decision tree based on Under-sampling and Attribute selection called UAMDT is proposed. First, Tomek link under-sampling and Ensemble Under-sampling are used for data processing, and many balanced subsets are obtained. Furthermore, single decision tree is constructed on each subset, the hybrid attribute measure of information gain and Gini index as attribute selection criteria are used and the optimal attribute as the split attribute of the root node of each single decision tree is selected, and finally all the single decision trees are integrated to build a multi-decision tree. In this paper, the experiments with multiple evaluation criteria on 10 imbalanced datasets are conducted to verify the effectiveness and feasibility of the proposed algorithm.

The Shortest-path exponent d_min and the Backbone exponent d_B are two important critical exponents for characterizing the universality class in phase transition in statistical physics. Due to the lack of exact solutions, the values of d_min and d_B can only be estimated by numerical methods, such as Monte Carlo simulation (MC). In this paper, by sampling the graph distance on the complete configuration and the size of the largest cluster on the bridge-free configuration, for the first time, the shortest-path and backbone exponents are estimated for the square-lattice explosive percolation with product rule as d_min=1.189(3) and d_B=1.546(5), respectively. The results provide important testing foundations for the future analytical investigations of the critical geometrical properties of the percolation models with non-trivial rules.

Based on the chemical-potential lattice Boltzmann multiphase flow model, a modified chemical-potential multiphase flow model is proposed by introducing “reduced variables”. The performance of the model is verified, and the results show that, compared with the original chemical potential model, the temperature range and the gas-liquid density ratio of the modified model are greatly extended, the stability is remarkably improved, and the effect of spurious current is effectively reduced. The model is in good agreement with the thermodynamic consistency, meets with Laplace’s law, and has practical and applicable prospects.

To overcome the problems of the existing road traffic sign detection methods, such as slow speed, large environmental impact and poor detection effect, a multi-target real-time detection method for road traffic signs based on Faster-RCNN is proposed. Firstly, the principle of Faster-RCNN target detection is analyzed in depth. Then, the Faster-RCNN network structure is optimized, and the appropriate pre-training model and network hyperparameters are selected. Finally, a set of comparative experiments are designed on the German Traffic Sign Detection dataset (GTSD), which prove the validity of the method. The detection time of single image is 0.4 s, and the accuracy rate is over 71%. The migration test is conducted on the Sweden traffic sign detection dataset (STSD). The method demonstrates a good generalization capability and provides a theoretical basis and technical support for the application of smart cars.

Aiming at the problem that the traditional distance-based outlier recognition method can not be directly and effectively applied to high-dimensional data and the recognition effect is affected by parameters, an adaptive high-dimensional outlier recognition method is proposed, which uses genetic algorithm. The optimized Gaussian Restricted Boltzmann machine nonlinearly maps high-dimensional data to low-dimensional space, and then performs outlier recognition in low-dimensional data space by adaptive outlier recognition. UCI high-dimensional data and high-dimensional data of traditional Chinese medicine are used to verify the experiment. The experimental results show that the adaptive high-dimensional outlier recognition method can adaptively and effectively identify outliers in high-dimensional data.

In view of the non-linear characteristics of the three-axis stabilized pan-tilt servo system, the anti-disturbance ability of PD control is poor, and the setting process of the active disturbance rejection control is time-consuming and laborious due to the large number of parameters. By using the global approximation ability and self-learning ability of BP neural network, a composite controller is composed of BP neural network and active disturbance rejection control. All the key parameters of active disturbance rejection control are self-tuned and optimized, which is applied to the three-axis stabilized pan-tilt servo system with Stribeck friction model. The simulation results show that the method is feasible and effective for parameter auto-tuning. Compared with the conventional ADRC with fixed parameters and PD control, it has higher control accuracy and stronger anti-disturbance ability, and has better application value for improving the performance of the stabilized platform.

It’s hard to get the contact force of screw turning accurately in a robot system. If a force sensor is used to detect the contact force, there are lots of problems, such as, it is hard to fix the sensor and noise of sensor or transmission delay of signal may affect the accuracy of system control. In this paper, instead of using the force sensor to detect the force. the torque value of the servo motor is used to compute the contact force between the head of screw driver and the work-piece. The expected locking torque is used as the threshold value in the control system design and the energy regulator is used in energy correction to balance the predictive force value and current calculation force value, in order that the control system of robot screwing system can be stable in passivity and active compliance. In the automatic screwing process of a precision plastic, the passive flexible mechanic and event based control method is designed to test the safety and feasibility of the presented hybrid acitive-passive force-position control method.

By using the Poincare ring domain theorem, Lyapunov-branch theorem, and constructing appropriate Poincare ring domain, it is shown that U(u,v,λ),V(u, v, λ) unanimously tends to 0 respectively as λ→0 in a bounded domain of the uov plane. When the parameter λ approaches the nonzero positive or positive infinity, the corresponding system has at least a stable or unstable limit cycle near the origin. For the two cases of the parameter λ approaches non-zero positive number and positive infinity, examples are given to verify the wide applicability of the conclusions.

This paper studies the existence of uniform attractors for a class of nonautonomous stochastic reaction-diffusion equations with white noise on unbounded domains. Firstly, with uniform estimation of the solutions, it is proved that the stochastic dynamical system corresponding to the original equation has a uniformly with respect to symbol space pullback absorbing set. Secondly, by asymptotic tail estimation, it is proved that the solution is uniformly pullback and asymptotically compact. The existence of uniform random attractor of the original system is obtained.

In this paper, a stochastic SIRS epidemic model with saturated incidence rate and double epidemic hypothesis is investigated. By constructing suitable Lyapunov function and applying Itô formula, the global existence and uniqueness of positive solution are proved, and the random basic reproductive number which determines disease extinction and persistence under certain conditions is obtained. The influence of disease is also discussed when the environment changes. The results show that the intensity of white noise suppresses the outbreak of the disease under certain conditions.The conclusions are simulated through the numerical method.

Signal separation approaches are often used for time series analysis because the extracted subcomponents of a complex signal can be helpful for analyzing its trend and periodicity. In recent years, with the emergence of adaptive signal separation algorithms, such as Null Space Pursuit (NSP) and Empirical Mode Decomposition (EMD), those methods have been widely used in the field of economic and financial time series analysis. The statistical data of loan/deposit and GDP have characteristics of developing trend and periodical fluctuation. Based on the NSP method, in this paper, a multi-scale study is conducted on the dynamic of regional economic disparity of 31 provinces in China during 2005-2017, and the volatility mechanism is investigated in the view of credit scale.