In recent years, deep learning models have been more and more widely used in various fields and have become a research hotspot for various tasks in artificial intelligence, but there are significant limitations in non-convex optimization problems. As a model training strategy for non-convex optimization, curriculum learning advocates that models learn in the order of easier to more difficult data, mimicking the basic idea of gradual human learning as they learn curriculum. This strategy has been widely used in the fields of computer vision, natural language processing, and reinforcement learning; it can effectively solve the non-convex optimization problem and improve the generalization ability and convergence speed of models. This paper first introduces the application of curriculum learning at three major levels: data, task, and model, and summarizes the evaluators designed using curriculum learning methods in various domains, including difficulty evaluators, training schedulers, and loss evaluators, which correspond to the three stages of difficulty evaluation, training schedule, and loss evaluation in the application of curriculum learning to model training. We also discuss how to choose an appropriate evaluation system and the differences between terms used in different types of research. Finally, we summarize five methods similar to curriculum learning in the field of machine learning and provide a summary and outlook of the curriculum learning evaluation system.

In recent years, significant progress has been made in the application of federated learning (FL) in various aspects of cyberspace security, such as intrusion detection, privacy protection, and anomaly detection. However, the robustness of federated learning in the face of malicious attacks (such us adversarial attacks, backdoor attacks, and poisoning attacks) is weak, and the unfair allocation of resources leads to slow convergence and inefficient communication efficiency regarding FL models. Additionally, the scarcity of malicious samples during FL model training and the heterogeneity of data result in a lack of personalization in FL models. These challenges pose significant obstacles to the application of federated learning in the field of cyberspace security. To address these issues, the introduction of meta-learning into federated learning has been proposed, resulting in the development of federated meta-learning models. These models aim to train personalized models for each client, reducing performance discrepancies across different clients and enhancing model fairness. In order to advance research on federated meta-learning and its applications in the field of cyberspace security, this paper first introduces the algorithms of federated meta-learning. Based on different usage principles, these algorithms are categorized into client-level personalization algorithms, network algorithms, prediction algorithms, and recommendation algorithms, and are thoroughly presented and analyzed. Subsequently, the paper divides current cyberspace security issues in the network domain into three branches: information content security, network security, and information system security. For each branch, the application research methods and achievements of federated meta-learning are elucidated and compared, highlighting the advantages and disadvantages of federated meta-learning in addressing different cyberspace security issues. Finally, the paper concludes with an outlook on the deep application of federated meta-learning in the field of cyberspace security.

Cyber physical systems (CPS), in the event of a cyber attack, can have a serious impact on the operating physical equipment. In order to improve the attack detection capability of CPS, an support vector machine (SVM) attacks detection model based on particle swarm optimization (PSO) is proposed. First, the box plot anomaly detection method is used to detect the characteristic variables, and the characteristic variables with abnormal distribution are discretized. Secondly, the number of attack samples was increased by the SMOTE method to solve the problem of data imbalance, and the linear combination of characteristic variables was performed on the high-dimensional CPS network traffic data using principal component analysis (PCA). Then, the penalty coefficient and the hyperparameter of the kernel function in the SVM model are optimized by the PSO algorithm. Finally, Experiments on attack detection of CPS network traffic data show that the proposed model can detect different types of attack data and has higher detection accuracy compared with general detection models.

With the rapid development of the Internet, malicious domain names pose more and more serious threats to many fields, such as network security and social security, and there have been many research results on malicious domain detection. This article proposes a malicious domain name detection model based on improved deep learning, which can combine the advantages of three different network models, convolutional neural network (CNN), temporal convolutional network (TCN), and long short-term memory network (LSTM) in malicious domain name detection, to obtain a better detection effect than that of the original single or two models. Experiments show that the effect of the improved deep learning model proposed in this article is better than that of the combined model of CNN and LSTM or the combined model of CNN and TCN, and the accuracy and regression rates reached 99.76% and 98.81%, respectively.

To address the accuracy issues in monocular depth estimation caused by insufficient feature extraction and inadequate context modeling, a multi-scale feature optimization model named EFDepth was proposed to improve prediction performance. This framework adopted an encoder–decoder structure: the encoder (EC-Net) was composed of MobileNetV3-E and ETFBlock, and its features were optimized through multi-scale dilated convolution; the decoder (LapFA-Net) combined the Laplacian pyramid and the FMA module to enhance cross-scale feature fusion and output accurate depth maps. Comparative experiments between EFDepth and algorithms including Lite-mono, Hr-depth, and Lapdepth were conducted on the KITTI datasets. The results show that, for the three error metrics—RMSE (Root Mean Square Error), AbsRel (Absolute Relative Error), and SqRel (Squared Relative Error)—EFDepth is 1.623, 0.030, and 0.445 lower than the average values of the comparison algorithms, respectively, and for the three accuracy metrics, it is 0.052, 0.023, and 0.011 higher than the average values of the comparison algorithms, respectively. Experimental results indicate that EFDepth outperforms the comparison methods in most metrics, providing an effective reference for monocular depth estimation and 3D reconstruction of complex scenes.