Narrow Results

The possible application of boosted neural network to particle classification in high energy physics is discussed. A two-dimensional toy model, where the boundary between signal and background is irregular but not overlapping, is constructed to show how boosting technique works with neural network. It is found that boosted neural network not only decreases the error rate of classification significantly but also increases the efficiency and signal–background ratio. Besides, boosted neural network can avoid the disadvantage aspects of single neural network design. The boosted neural network is also applied to the classification of quark- and gluon-jet samples from Monte Carlo e⁺e^- collisions, where the two samples show significant overlapping. The performance of boosting technique for the two different boundary cases — with and without overlapping is discussed.

Privacy and security are the most concerning topics while using cloud-based applications. Malware detection in cloud applications is important in identifying application malware activity. So, a novel Goat-based Recurrent Forensic Mechanism (GbRFM) is used to detect the attack and provide the attack type in cloud-based applications. At first, the dataset is pre-processed in the hidden phase, and the errorless features are extracted. The proposed model also trains the output of the hidden layer to identify and classify the malware. The wild goat algorithm enhances the identification rate by accurately detecting the attack. Using the NSL-KDD data, the preset research was verified, and the outcomes were evaluated. The performance assessment indicates that the developed model gained a 99.26% accuracy rate for the NSL-KDD dataset. Moreover, to validate the efficiency of the proposed model, the outcomes are compared with other techniques. The comparison analysis proved that the proposed model attained better results.

As blockchain technology and smart contracts develop, computer technology is constantly integrating with smart chemical plants. Due to the continuous development of intelligent chemical plants, their systems have gradually become large and dispersed, posing a threat to safety management. In order to improve the performance of intelligent security management systems, the study first explores the principles of blockchain and smart contract technology, and then combined with the requirements of intelligent chemical plant security management systems, designs an intelligent security management system based on blockchain and smart contract technology. The experimental results showed that compared to systems without smart contract support, the communication success rate between nodes was lower. The error rates of blockchain-based encryption systems, deep learning-based encryption systems and improved data encryption systems proposed in the study were 0.22, 0.07 and 0.09, respectively. The packet loss rates were 0.13, 0.04 and 0.05, respectively. The lower the bit error rate and packet loss rate of the encryption system, the clearer the illegal eavesdropping information. The experimental results indicate that the intelligent security management system designed in this study has good encryption performance and a higher communication success rate. The results have certain reference value in security management application in intelligent chemical plants.

Semi-quantum key distribution (SQKD) can share secret keys by using less quantum resource than its fully quantum counterparts, and this likely makes SQKD become more practical and realizable. In this paper, we present a new SQKD protocol by introducing the idea of B92 protocol in fully quantum cryptography into SQKD. In this protocol, the sender Alice just sends one quantum state to the classical Bob and Bob just prepares a fixed state in the preparation process. It is much simpler than the existing SQKD and potentially much easier to be implemented. It can be seen as a semi-quantum version of B92 protocol, compared to the protocol BKM07 as the semi-quantum version of BB84 in fully quantum cryptography. We verify that it is more efficient than the existing single-state SQKD protocols by introducing an efficiency parameter. Specifically, we prove it is secure against a restricted collective attack by computing a lower bound of the key rate in the asymptotic scenario. Then we can find a threshold value of errors such that for all error rates less than this value, the secure key can be definitely established between the legitimate users definitely. We make an illustration of how to compute the threshold value in case the reverse channel is a depolarizing one with parameter $p$. Though the threshold value is a little smaller than those of some existing SQKD protocols, it can be comparable to the B92 protocol in fully quantum cryptography.

In this paper, BP neural network model is used to predict the compression strength and thermal conductivity of the foamed concrete. The experimental data were divided into training dataset and control dataset. On the training dataset, the proposed BP neural network model was applied. The fitted model was obtained by tuning the parameters of mixing proportion with error rate controlled at pre-defined level. The prediction accuracy of the model was verified by comparing the results of the fitted model on the control dataset with true values. The results show that the predicted error rate is less than 8%, indicating that BP neural network is capable of predicting the experimental data accurately.