With the rapid development of networks, users are increasingly seeking richer and high-quality content experience, and there is an urgent need to develop efficient content caching strategies to improve the content distribution efficiency of caching. Therefore, it will be an effective solution to combine content popularity prediction based on machine learning (ML) and content caching to enable the network to predict and analyze popular content. However, the data sets which contain users’ private data cause the risk of privacy leakage. In this paper, to address this challenge, we propose a privacy-preserving algorithm based on federated learning (FL) and long short-term memory (LSTM), which is referred to as FL-LSTM, to predict content popularity. Simulation results demonstrate that the performance of the proposed algorithm is close to the centralized LSTM and better than other benchmark algorithms in terms of privacy protection. Meanwhile, the caching policy in this paper raises about 14.3% of the content hit rate.

The free space optical communication plays an important role in space-terrestrial integrated network due to its advantages including great improvement of data rate performance, low cost, security enhancement when compared with conventional radio frequency (RF) technology. Meanwhile, CubeSats become popular in low earth orbit (LEO) network because of the low cost, fast response and the possibility of constituting constellations and formations to execute missions that a single large satellite cannot do. However, it is a difficult task to build an optical communication link between the CubeSats. In this paper, the cutting-edge laser technology progress on the CubeSats is reviewed. The characters of laser link on the CubeSat and the key techniques in the laser communication terminal (LCT) design are demonstrated.

The Very Fast Decision Tree (VFDT) algorithm is a classification algorithm for data streams. When processing large amounts of data, VFDT requires less time than traditional decision tree algorithms. However, when training samples become fewer, the label values of VFDT leaf nodes will have more errors, and the classification ability of single VFDT decision tree is limited. The Random Forest algorithm is a combinational classifier with high prediction accuracy and noise-tolerant ability. It is constituted by multiple decision trees and can make up for the shortage of single decision tree. In this paper, in order to improve the classification accuracy on data streams, the Random Forest algorithm is integrated into the process of tree building of the VFDT algorithm, and a new Random Forest Based Very Fast Decision Tree algorithm named RFVFDT is designed. The RFVFDT algorithm adopts the decision tree building criterion of a Random Forest classifier, and improves Random Forest algorithm with sliding window to meet the unboundedness of data streams and avoid process delay and data loss. Experimental results of the classification of KDD CUP data sets show that the classification accuracy of RFVFDT algorithm is higher than that of VFDT. The less the samples are, the more obvious the advantage is. RFVFDT is fast when running in the multi-thread mode.