With the rapid development of semiconductor technology, traditional equation-based modeling faces challenges in accuracy and development time. To overcome these limitations, neural network (NN)-based modeling methods have been proposed. However, the NN-based compact model encounters two major issues. Firstly, it exhibits unphysical behaviors such as un-smoothness and non-monotonicity, which hinder its practical use. Secondly, finding an appropriate NN structure with high accuracy requires expertise and is time-consuming. In this paper, we propose an Automatic Physical-Informed Neural Network (AutoPINN) generation framework to solve these challenges. The framework consists of two parts: the Physics-Informed Neural Network (PINN) and the two-step Automatic Neural Network (AutoNN). The PINN is introduced to resolve unphysical issues by incorporating physical information. The AutoNN assists the PINN in automatically determining an optimal structure without human involvement. We evaluate the proposed AutoPINN framework on the gate-all-around transistor device. The results demonstrate that AutoPINN achieves an error of less than 0.05%. The generalization of our NN is promising, as validated by the test error and the loss landscape. The results demonstrate smoothness in high-order derivatives, and the monotonicity can be well-preserved. We believe that this work has the potential to accelerate the development and simulation process of emerging devices.

Due to the advantages and limitations of the two kinds of vulnerability mining methods of static and dynamic analysis of android applications, the paper proposes a method of Android application vulnerability mining based on dynamic and static combination. Firstly, the static analysis method is used to obtain the basic vulnerability analysis results of the application, and then the input test case of dynamic analysis is constructed on this basis. The fuzzy input test is carried out in the real machine environment, and the application security vulnerability is verified with the taint analysis technology, and finally the application vulnerability report is obtained. Experimental results show that compared with static analysis results, the method can significantly improve the accuracy of vulnerability mining.

Traditional network intrusion detection algorithms tend to lack learning in a small number of classes due to data imbalance. In reality, intrusion detection systems pay more attention to the detection accuracy of a small number of classes, that is, attack samples. In order to improve the detection accuracy of intrusion detection system, a network intrusion detection method based on deep convolution generative adversarial networks (DCGAN) was proposed. Firstly, the sample data of network intrusion is preprocessed, and the character data set is replaced by image data. Then, DCGAN is used to train and test the sample data. Both the generator and the discriminator are constructed by CNN. The generator is used to construct attack samples, balance the number of training samples, and solve the over fitting problem caused by insufficient training samples. Finally, the trained discriminator is used to test the classification accuracy of samples. Experimental results show that, compared with the traditional algorithm, the proposed algorithm can not only balance the detection accuracy of various types of samples, but also has higher detection accuracy for attack samples.

With the development of computer technology and communication technology, computer network will increasingly become an important means of information exchange, and permeate into every field of social life. However, the network has the potential threat and the reality existence each kind of security question, therefore we must take the strong security policy to ensure the network security. The purpose of this paper is to study the network computer security hidden trouble and vulnerability mining technology. In this paper, the types of security hidden dangers are analyzed, and the vulnerability detection technology Fuzzing technology is deeply studied. Then, the inspection and test time is analyzed for the existing vulnerability detection tools. The experimental results prove that vulnerability detection technology can protect network security with high efficiency of vulnerability detection. In this paper, three vulnerability detection tools WS Fuzzer, Web Fuzz and Webvul Scan were used to analyze the detection time of open source system, personal blog, shopping mall and forum. The average detection time was 1.9s, 8.7s, 20.5s and 59.7s, respectively. It can be seen that the vulnerability mining technology has a certain practical role.

This article proposes a method to generate accurate and concise analytic equations for device compact modeling using Physics-Assisted Kolmogorov–Arnold Networks (PKAN). The equations are directly extracted from the trained neural network architecture. PKAN uses variable activation functions informed by prior physical knowledge to model device behaviors. Similarity constraints map these trained activation functions to mathematical symbols. Sparsification techniques simplify the network structure, producing concise and explicit equations. This article also presents four approaches for physics-assisted device modeling using PKAN: (1) generating entire continuous equations without human intervention, (2) applying correlation factors to existing models without requiring knowledge of internal physical mechanisms, (3) revising specific parts of existing models, and (4) automatically extending existing models. Experimental results show that PKAN demonstrates significant accuracy improvements, achieving error reductions of 91.8%, 91.5%, 66.2%, and 83.7% for corresponding experiments, respectively. These findings demonstrate PKAN’s potential for various device modeling applications. By combining the precision of neural networks with the clarity of symbolic representation, PKAN offers a powerful tool for device modeling applications.