Liqiang Zhang

The continuous growth of the global population has led to an increasing demand for food, yet the annual food losses caused by pests are immeasurable. Therefore, accurately identifying crop pests and adopting effective control strategies is of great practical significance. In recent years, intelligent crop pest classification models based on deep learning have made significant progress, which can be primarily divided into two categories: convolutional neural network (CNN)-based and Transformer-based. However, the ability of CNN is limited in modeling the relationship between long-distance regions in images; the capability of Transformer is insufficient in capturing the local discriminative features of crop pests, which leads to unsatisfactory performance of the above two intelligent crop pest classification models. To address the above issues, this paper proposes a new crop pest classification model (CT-PestNet). First, CT-PestNet combines CNN and Transformer to fully leverage the former’s strengths in local detail extraction and the latter’s capabilities in modeling global dependencies. Second, a cross-scale feature distillation enhancement module (CFDE) is designed to respond to the challenges of diverse pest morphologies (with varying sizes and shapes) and complex backgrounds. The CFDE employs the “distillation-then-enhancement” strategy to enhance the model’s perception of small or diverse pests, thereby improving its classification accuracy and robustness in complex backgrounds. Finally, a local token interaction module (LTIR) is constructed to overcome the insufficient modeling of adjacent token features in the Transformer encoder. The LTIR can achieve deep discriminative feature mining by enhancing the feature interaction among adjacent tokens, thereby improving the model’s processing ability and recognition performance for complex images. On the IP102 and D0 datasets, CT-PestNet achieves accuracies of 77.32% and 99.73%, respectively; especially on the more challenging IP102 dataset, it outperforms the best CNN-based model by 3.62% and the best Transformer-based model by 1.32%, fully proving its effectiveness.

The reliability and service life of steel products are largely dependent on the quality of the steel strip, so it is crucial to accurately identify its surface defects. Steel strip surface defect data collected from actual industrial scenarios is often limited and mostly unlabeled, so researchers often use Generative Adversarial Networks (GANs) for data augmentation while introducing semi-supervised learning to improve the performance of deep learning-based steel strip surface defect classification models. However, the samples generated by GANs often lack sufficient diversity and fail to simulate the actual distribution of the real defect data. In addition, strip surface defect data often exhibits highly similar texture features at the fine-grained level, which significantly degrades the performance of models. Therefore, a new semi-supervised strip steel surface defect classification model is proposed. Firstly, a data augmentation strategy based on CycleGAN is developed to generate unlabeled samples with more diversity and fit real industrial scenarios, thereby improving the model’s generalization ability. Secondly, a feature extraction and classification network based on ResNet-50 is designed to enhance the model’s capability to capture fine-grained features, thus improving the ability to distinguish between different defect categories. Finally, extensive experiments were conducted on the NEU-CLS dataset. The results showed that when trained with the 10% labeled data and the 90% unlabeled data, the proposed model achieved accuracy, precision, and recall rates of 92.91%, 89.64%, and 84.81% in this dataset, indicating competitive performance compared with multiple existing semi-supervised models.

This paper has determined the relevance of the issue related to improving the accuracy of the results of mathematical modeling of the software security testing process. The fuzzy GERT-modeling methods have been analyzed. The necessity and possibility of improving the accuracy of the results of mathematical formalization of the process of studying software vulnerabilities under the conditions of fuzziness of input and intermediate data have been determined. To this end, based on the mathematical apparatus of fuzzy network modeling, a fuzzy GERT model has been built for investigating software vulnerabilities. A distinctive feature of this model is to take into consideration the probabilistic characteristics of transitions from state to state along with time characteristics. As part of the simulation, the following stages of the study were performed. To schematically describe the procedures for studying software vulnerabilities, a structural model of this process has been constructed. A "reference GERT model" has been developed for investigating software vulnerabilities. The process was described in the form of a standard GERT network. The algorithm of equivalent transformations of the GERT network has been improved, which differs from known ones by considering the capabilities of the extended range of typical structures of parallel branches between neighboring nodes. Analytical expressions are presented to calculate the average time spent in the branches and the probability of successful completion of studies in each node. The calculation of these probabilistic-temporal characteristics has been carried out in accordance with data on the simplified equivalent fuzzy GERT network for the process of investigating software vulnerabilities. Comparative studies were conducted to confirm the accuracy and reliability of the results obtained. The results of the experiment showed that in comparison with the reference model, the fuzziness of the input characteristic of the time of conducting studies of software vulnerabilities was reduced, which made it possible to improve the accuracy of the simulation results.

This paper reports an analysis of the software (SW) safety testing techniques, as well as the models and methods for identifying vulnerabilities. An issue has been revealed related to the reasoned selection of modeling approaches at different stages of the software safety testing process and the identification of its vulnerabilities, which reduces the accuracy of the modeling results obtained. Two steps in the process of identifying software vulnerabilities have been identified. A mathematical model has been built for the process of preparing security testing, which differs from the known ones by a theoretically sound choice of the moment-generating functions when describing transitions from state to state. In addition, the mathematical model takes into consideration the capabilities and risks of the source code verification phase for cryptographic and other ways to protect data. These features generally improve the accuracy of modeling results and reduce input uncertainty in the second phase of software safety testing. An advanced security compliance algorithm has been developed, with a distinctive feature of the selection of laws and distribution parameters that describe individual state-to-state transitions for individual branches of Graphical Evaluation and Review Technique networks (GERT-networks). A GERT-network has been developed to prepare for security testing. A GERT-network for the process of checking the source code for cryptographic and other data protection methods has been developed. A graphic-analytical GERT model for the first phase of software safety testing has been developed. The expressions reported in this paper could be used to devise preliminary recommendations and possible ways to improve the effectiveness of software safety testing algorithms

The subject of the research is methods and algorithms for developing the GERT networks that provide a hidden transition along the branches of the control logic graph and the ability to encode digital watermarks in the context of existing threats to the licensing security of software systems. The aim of the work is to develop a system of a software product licensed security based on hidden transition algorithms in GERT networks that carry the properties of digital watermarks. The article solves the following goal: development of a model of a software licensed security system based on the developed algorithms using GERT networks. Methods of mathematical modeling, numerical experiment, complexity theory and cryptography are used. The following results were obtained: on the basis of the analysis of existing models for ensuring software licensed security, the main requirements for the algorithms synthesized within the framework of the developed model, as well as for the machine implementation of the algorithm, were identified. Based on the investigation of existing watermark systems, methods of attack on them, as well as the requirements put forward for the developed model, a licensing security algorithm was formed based on the watermark system. Conclusions: for the first time, algorithms for safe transition in GERT-networks were developed, which are used as a graph of the control logic of a software product. This logic is implemented depending on the identification or serial number; a model of a licensed security system has been developed, which has not only empirical, but also theoretical substantiation of resistance to attacks by an intruder.