Federated Learning （FL） allows users to share and interact with multiple parties without directly uploading the original data， effectively reducing the risk of privacy leaks. However， existing research suggests that the adversary can still reconstruct raw data through shared gradient information. To further protect the privacy of federated learning， a deep shadow defense scheme of federated learning based on Generative Adversarial Network （GAN） was proposed. The original real data distribution features were learned by GAN and replaceable shadow data was generated. Then， the original model trained on real data was replaced by a shadow model trained on shadow data and was not directly accessible to the adversary. Finally， the real gradient was replaced by the shadow gradient generated by the shadow data in the shadow model and was not accessible to the adversary. Experiments were conducted on CIFAR10 and CIFAR100 datasets for comparison of the proposed scheme with the five defense schemes of adding noise， gradient clipping， gradient compression， representation perturbation and local regularization and sparsification. On CIFAR10 dataset， the Mean Square Error （MSE） and the Feature Mean Square Error （FMSE） of the proposed scheme were 1.18-5.34 and 4.46-1.03×10⁷ times， and the Peak Signal-to-Noise Ratio （PSNR） of the proposed scheme was 49.9%-90.8%. On CIFAR100 dataset， the MSE and the FMSE of the proposed scheme were 1.04-1.06 and 5.93-4.24×10³ times， and the PSNR of the proposed scheme was 96.0%-97.6%. Compared with the deep shadow defense method， the proposed scheme takes into account the actual attack capability of the adversary and the problems in shadow model training， and designs threat models and shadow model generation algorithms. It performs better in theory analysis and experiment result that of the comparsion schemes， and it can effectively reduce the risk of federated learning privacy leaks while ensuring accuracy.

In intelligent computing services， data analysis and processing are provided for the service consumer by the service provider through Internet， and a learning model is established to complete intelligent computing function. Due to the lack of effective communication channels between service providers and service consumers， as well as the fuzzy and messy requirement descriptions of the service consumer feedback， there is a lack of a unified service requirement acquisition method to effectively analyze， organize and regulate the continuously changing requirement of users， which leads to the failure of intelligent computing services to make a rapid improvement according to the user’s requirements. Aiming at the problems of continuity and uncertainty of requirement changes in service development， a requirement acquisition method for intelligent computing services was proposed. The application feedback and questions of intelligent computing services were firstly obtained from Stack Overflow question and answer forum. Then， the knowledge classification and prioritization were performed on them by using different learning models （including Support Vector Machine （SVM）， naive Bayes and TextCNN） according to the types of requirements concerned by the service consumer. Finally， a customized service requirement template was used to describe the requirements of intelligent computing services.

For analyzing the effect of the call distance on the detecting probability in call magnetic anomaly searching submarine, the model for calculating the submarine distribution probability was deduced, and then the relation between the call distance and the call magnetic anomaly searching submarine probability was established. At last, some calculation results were given out for some typical cases. The results show the call magnetic anomaly searching submarine probability descends rapidly with the call distance increasing, just only in near call distance, small initial distribution radius and low velocity, the call magnetic anomaly searching submarine has a high detecting probability. This shows that the call distance has a serious effect on the detecting probability in call magnetic anomaly searching submarine, and the magnetic anomaly detecting is not fit for searching submarine for far call distance.

The application of traditional software prediction model has poor generalized performance. This paper put forward a software reliability prediction model based on Learning Vector Quantization (LVQ) neural network. First, this paper analyzed the structure characteristics of LVQ neural network and its relation with software reliability prediction. Then the network was used to predict the software reliability. In the end, the authors confirmed the algorithm through multiple simulation experiments under the Matlab environment and the data from Metrics Data Program (MDP) database of National Aeronautics and Space Administration （NASA） of USA. The experimental results indicate that the method is feasible and has a higher prediction precision than the traditional software prediction method.

In literature, there are very few discussions on the change of performance of source camera classification algorithms when test images are subjected to minor image processing. Using Support Vector Machines (SVM), this paper analyzed the performance and robustness of source camera classification algorithms. It compared the detection accuracy for unprocessed images with that for processed images, and investigated the robustness of different types of image features. Since pattern classification-based algorithms often need to reduce the number of image features for computational efficiency, this paper also discussed the performance of camera classification algorithms using the image feature subsets. The impact of using these subsets on the robustness of camera classification algorithms was explored as well.