To solve the problem of inaccurate localization of model which is caused by inaccurate manual labeling in 3D point cloud training data, a novel robust 3D object detection method based on localization uncertainty was proposed. Firstly, with the 3D voxel grid-based Sparsely Embedded CONvolutional Detection (SECOND) network as basic network, the prediction of localization uncertainty was added based on Region Proposal Network (RPN). Then, during the training process, the localization uncertainty was modeled by using Gaussian and Laplace distribution models, and the localization loss function was redefined. Finally, during the prediction process, the threshold filtering and Non-Maximum Suppression (NMS) methods were performed to filter candidate objects based on the object confidence which was consisted of the localization uncertainty and classification confidence. Experimental results on the KITTI 3D object detection dataset show that compared with SECOND network, the proposed algorithm has the detection accuracy improved by 0.5 percentage points on car category at moderate level. The detection accuracy of the proposed algorithm is 3.1 percentage points higher than that of SECOND network with adding disturbance simulation noise to the training data in the best case. The proposed algorithm improves the accuracy of 3D object detection, which reduces false detection and improves the accuracy of 3D bounding boxes, and is more robust to noisy data.

The social recommendation system can alleviate the data sparsity and cold start problems in the recommendation system through the users' social attribute information, thereby improving the accuracy of the recommendation system. However, most social recommendation methods mainly aim at the single social network or linearly superimpose multiple social networks, making it difficult for the users' social attributes to fully participate in the calculation, so the accuracy of recommendation is limited. To solve this problem, a multi-network embedded graph neural network model was proposed to implement the recommendation in complex multidimensional social networks. In the model, a unified method was built to fuse the multidimensional complex networks composed of user-item, user-user and other relationships. Different types of multi-neighbors were aggregated to attribute to the node generation through attention mechanism, and multiple graph neural networks were combined to construct a graph neural network recommendation framework under multidimensional social relationships. In the proposed method, the entities in the recommendation system and their relationships were reflected by the topology structure, and the relevant information was calculated and updated continuously on the graph directly. It can be seen that the method is inductive, and avoids the problem of incomplete information utilization in traditional recommendation methods effectively. By comparing with related social recommendation algorithms, the experimental results show that the recommendation accuracy indicators such as Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) of the proposed method are improved, and the method even has good accuracy on sparse data.

To solve the multi-label classification problem that a target belongs to multiple classes, a new multi-label classification algorithm based on floating threshold classifiers combination was proposed. Firstly, the theory and error estimation of the AdaBoost algorithm with floating threshold (AdaBoost.FT) were analyzed and discussed, and it was proved that AdaBoost.FT algorithm could overcome the defect of unstabitily when the fixed segmentation threshold classifier was used to classify the points near classifying boundary, the classification accuracy of single-label classification algorithm was improved. And then, the Binary Relevance (BR) method was introduced to apply AdaBoost.FT algorithm into multi-label classification problem, and the multi-label classification algorithm based on floating threshold classifiers combination was presented, namely multi-label AdaBoost.FT. The experimental results show that the average precision of multi-label AdaBoost. FT outperforms the other three multi-label algorithms, AdaBoost.MH (multiclass, multi-label version of AdaBoost based on Hamming loss), ML-kNN (Multi-Label k-Nearest Neighbor), RankSVM (Ranking Support Vector Machine) about 4%, 8%, 11% respectively in Emotions dataset, and is just little worse than RankSVM about 3%, 1% respectively in Scene and Yeast datasets. The experimental analyses show that multi-label AdaBoost. FT can obtain the better classification results in the datasets which have small number of labels or whose different labels are irrelevant.

It is called labels matching problem when two labels of an instance come from two labelsets respectively in multi-label classification, however there is no any specific algorithm for solving such problem. Although the labels matching problem could be solved by tranditional multi-label classification algorithms, but this problem has its own particularity. After analyzing the labels matching problem, a new labels matching algorithm based on pairwise labelsets was proposed using adaptive method, which considered the real Adaptive Boosting (real AdaBoost) and the global optimization idea. This algorithm could learn the rule of labels matching well and complete matching. The experimental results show that, compared with the traditional algorithms, the new algorithm can not only reduce searching scope of the labels space, but also decrease the minimum learning error as the number of weak classifiers increases, and make the classification more accurate and faster.

In view of the problems that posture recognition based on vision requires a lot on environment and has low anti-interference capacity, a posture recognition method based on predefined bone was proposed. The algorithm detected human body by combining Kinect multi-scale depth and gradient information. And it recognized every part of body based on random forest which used positive and negative samples, built the body posture vector. According to the posture category, optimal separating hyperplane and kernel function were built by using improved support vector machine to classify postures. The experimental results show that the recognition rate of this scheme is 94.3%, and it has good real-time performance, strong anti-interference, good robustness, etc.

According to the shape features of mechanical design images, an asynchronous retrieval method based on weighted distance was proposed. The algorithm firstly got preliminary results from the image database by using the circumcircle distance feature, and then calculated the weighted distances between the input image and the preliminary results, by considering both the formal output positions and the Hu invariant moments feature. The experiments show that compared with the traditional methods, the proposed method gets higher precision and recall ratio.

Concerning the complexity of network security management and the absence of self-adaptation on situation awareness process, a Network Security Situation Awareness Model (NSSAM) based on autonomic computing was proposed. The situation extraction was analyzed in real-time by an autonomic feedback law. From the perspectives of attack and defense, a multi-level and multi-angle network security situation assessment model employing Analytic Hierarchy Process (AHP) was established according to the extracted situation information. The model of future network security situation prediction adopting improved genetic neural network was built on the basis of the past and current network security situation. Test results show that NSSAM with autonomic feedback mechanism can effectively enhance self-adaptation of the system.

To analyze the abnormal behavior of program induced by software defects, a characterization method of program behavior was proposed firstly, and then the baseline behavior and fault behavior of program got defined and formally described. A quantitative approach to represent the fault behavior of program was proposed afterwards. Furthermore, a Program Fault Behavior Analysis (PFBA) was delivered and implemented, which selected system-call as state granularity of program behavior. Based on specific embedded benchmarks, the experiment was followed through with fault injection method to obtain early-described indices of fault behavior. The experimental results show that there exists a difference among program behaviors under each individual fault type. Based on an in-depth analysis, it is demonstrated that the diversity of fault behaviors is induced by algorithm implementations and structural characteristics of embedded program themselves. Therefore, the analysis of fault behavior presented here can reveal the characteristics of embedded program response behavior under specific software defects, as well as providing important feedback to the process of program development.

To realize the building and controlling of workflow patterns involving multiple instances, a solution was proposed from the perspective of network partition. The implementing method was discussed based on RTWD net proposed by HU Fei-hu, et al. in Patent China 201010114083.9. First, the sub-workflows involving multiple instances should be divided into a subnet. Then the related parameters of multiple instances were defined, and multiple instances were controlled based on it. The paper discussed the controlling of sequential, synchronous and asynchronous parallel workflow patterns involving multiple instances based on the method. Because the divided subnet keeps consistent with the definition of workflow model, multiple instances can be scheduled by original workflow engine, which simplifies the realization of multiple instance patterns.