Map is an effective tool to recognize real space. Cyberspace also needs abstract, symbolic and numerical maps for visualization and cognition. Based on a comprehensive review of existing theories and methods on cyberspace map and from a perspective of visual representation of cyberspace elements and cyberspace knowledge, the current study proposes a concept and connotation of the cyberspace geographic map by synthesizing the concept of“tupu” in both computer science and geography.. Besides, this study introduces the methods and key technologies for cyberspace geographic map construction, and outlooks the application fields from three aspects: cyberspace resource management, cyberspace behavior cognition, and comprehensive analysis of cyberspace events. Mapping the geographical map of cyberspace can associate cyberspace with real space and clearly express the structure of cyberspace. It is a necessary way to realize the cognition of cyberspace behavior and improve the capability of cybersecurity.

This paper comes up with a fast-training model based on random walk to address the problems of slow training speed for representation learning of large-scale cybersecurity knowledge graph and lack of relational representation of head and tail entities,. The model first performs an initial training representation of the entities of the overall knowledge graph by random walk under relational paths, then, a subject-object embedding is designed to learn the syntactical meaning of the relations in the knowledge graph by combining the relation-specific subject embedding with the relation-specific object embedding. Finally, fast training of the knowledge graph is again assisted by random wandering under relational paths. In this paper, extensive experiments are conducted on several datasets and the results are compared with those using several existing models. The results show that the model proposed in this paper can shorten the training time by 1/3 and improve representation by about 3%, effectively improving the representation learning effect while speeding up the training speed of knowledge graph representation learning.

With the development of information technology, space-based information system has expanded the cyberspace from ground to natural space. As a result, cyberspace is becoming more and more complex. And the problem of network security is also increasingly serious. Therefore, how to display and analyze the situation of space-based information systems in cyberspace comprehensively and deeply has become an urgent problem to be solved. This research analyzes the current situation of space-based information network and summarizes the characteristics of cyberspace. For space-based information systems and based upon the existing theories and methods of cyberspace visualization, this research explores the mapping relationship between cyberspace and geographic space and proposes the content and technical path of cyberspace visualization from three aspects: information elements, topological relations, and security behaviors. Moreover, it constructs a cyberspace geographic graph. This research may provide technical support for cognizing cyberspace situation and maintaining network security of space-based information systems.

The recent advances made by graph-based deep learning have demonstrated its great potential in processing non-Euclidean structured data, and a large number of research efforts have attempted to apply graph embeddings or graph neural networks to vulnerability detection. This survey systematically investigates the vulnerability detection based on graph deep learning. Firstly, we summarize the four main stages of the vulnerability detection process, including data set, graph data preparation, graph deep learning model construction, and result evaluation. Then, starting from the effectiveness of graph-based deep learning vulnerability detection, we respectively expound the research results based on code patterns, code similarity and specific application scenarios. Finally, by sorting out and summarizing the existing research works, we analyze the challenges and foresee the trends in this research field.

In view of the increasingly serious problem of network security, geographical space features are added into the prediction process to realize spatio-temporal prediction of network space elements in this study. Considering the research status that network data are often rarely combined with geospatial characteristics in the prediction process of network security elements, network vulnerability detection data with geospatial characteristics are also selected to construct the spatio-temporal data set of network vulnerabilities. By constructing a spatio-temporal graph convolution model combining graph convolution and gated time convolution, the development of network vulnerability situation can be predicted. ARIMA and LSTM temporal prediction models are selected for comparative experiments, and the proposed network vulnerability spatio-temporal graph convolution prediction model shows better prediction effect under MAE, RMSE and MAPE evaluation criteria.

To address the shortage of spatio-temporal feature extraction in network security situation prediction, a local spatio-temporal convolution-based network vulnerability prediction method, namely the local spatio-temporal graph convolutional network model, is proposed, and HA, LSTM, SVR and STGCN models are selected for comparison experiments on network vulnerability data. Experimental results show that the model proposed in this paper can effectively improve the accuracy in predicting the time and location of vulnerabilities as well as the type of network vulnerabilities.