Aiming at the problem that the cross-modality interaction and the impact of the contribution of each modality on the final sentiment classification results are not considered in multimodal sentiment analysis of video, a multimodal sentiment analysis model of Attention Mechanism based feature Fusion-Bidirectional Gated Recurrent Unit (AMF-BiGRU) was proposed. Firstly, Bidirectional Gated Recurrent Unit (BiGRU) was used to consider the interdependence between utterances in each modality and obtain the internal information of each modality. Secondly, through the cross-modality attention interaction network layer, the internal information of the modalities were combined with the interaction between modalities. Thirdly, an attention mechanism was introduced to determine the attention weight of each modality, and the features of the modalities were effectively fused together. Finally, the sentiment classification results were obtained through the fully connected layer and softmax layer. Experiments were conducted on open CMU-MOSI (CMU Multimodal Opinion-level Sentiment Intensity) and CMU-MOSEI (CMU Multimodal Opinion Sentiment and Emotion Intensity) datasets. The experimental results show that compared with traditional multimodal sentiment analysis methods (such as Multi-Attention Recurrent Network (MARN)), the AMF-BiGRU model has the accuracy and F1-Score on CMU-MOSI dataset improved by 6.01% and 6.52% respectively, and the accuracy and F1-Score on CMU-MOSEI dataset improved by 2.72% and 2.30% respectively. AMF-BiGRU model can effectively improve the performance of multimodal sentiment classification.

With the development of virtualization technology, co-resident attack becomes a common means to steal sensitive information from users. Aiming at the hysteresis of existing virtual machine dynamic migration method reacting to co-resident attacks, a virtual network function migration strategy based on security threat prediction in the context of 5G network slicing was proposed. Firstly, network slicing operation security was modeled based on Hidden Markov Model (HMM), and the network security threats were predicted by multi-source heterogeneous data. Then according to the security prediction results, the migration cost was minimized by adopting the corresponding virtual network function migration strategy. Simulation experimental results show that the proposed strategy can effectively predict the security threats and effectively reduce the migration overhead and information leakage time by using HMM, which has a better defense effect against co-resident attack.

Aiming at the problem that pilot overhead required by downlink channel estimation of FDD (Frequency-Division Duplexing) massive MIMO (Multiple-Input Multiple-Output) was unaffordable, a pseudo-random pilot optimization scheme based on Compressive Sensing (CS) techniques with non-orthogonal pilot at the base station and the objective to minimize the cross correlation of the measurement matrix was proposed firstly. Then, a crossover and mutation judgment mechanism and an inner loop and outer loop mechanism were introduced to ensure the optimization of pilot sequence. Secondly, a Channel State Information (CSI) estimation algorithm based on CS techniques by utilizing the spatially common sparsity and temporal correlation in wireless MIMO channels was presented. Matrix estimation is performed by using LMMSE (Linear Minimum Mean Square Error) algorithm to accurately obtain CSI. Analysis and simulation results show that compared with random search pilot optimization scheme, location-based optimization scheme, local common support algorithm, Adaptive Structured Subspace Pursuit (ASSP) algorithm, Orthogonal Matching Pursuit (OMP) algorithm and Stepwise Orthogonal Matching Pursuit (StOMP) algorithm, the proposed algorithm can significantly achieve good channel estimation performance in the case of low pilot overhead ratio and low Signal-to-Noise Ratio (SNR).

Aiming at the problem of high energy consumption caused by the increase of base station density in heterogeneous dense network, an energy-efficient method for micro base station deployment in heterogeneous networks was proposed. Firstly, the feasibility of micro base station positions was considered to mitigate the effects of environmental conditions. Then the optimization target value was weighed under different user distribution probability to enhance adaptability for different user distribution scenarios. Finally, an energy-efficient deployment algorithm for micro base stations was proposed by jointly optimizing the number, deployment position and power configuration of micro base stations. Simulation results show that the proposed method improves energy efficiency by up to 26% compared with the scheme which only optimizes the number and location of micro base stations. The experimental results demonstrate that the combined optimization method can improve the energy efficiency of the system compared with the deployment method without considering the power factor, and verifies the influence of the micro base station power on the energy efficiency of heterogeneous network.

In order to deal with the cache pollution attacks in Content Centric Networking (CCN), a defense scheme based on cache diversification was proposed. To reduce the attack scope, the in-network content services were divided into three categories and different cache strategies were used for different services. For private and real-time services, contents were directly delivered without being cached; for streaming media services, contents were pushed to be cached in the edge of network according to probablity; for document services, the priority was caching contents in the upstream, then pushing them to the downstream. Then different defense methods were configured on different nodes. For the edge nodes, attacks were detected by observing the request probability variation of different contents; for the upstream nodes, contents with low request rate were ruled out from the cache space by setting filter rules. The simulation results show that the network average hit ratio under service diversification mechanism is 17.3% higher than that under CCN with traditional caching strategies.The proposed scheme can effectively improve the defense capability of the network for the cache pollution attack.

A method of object detection with few samples based on two-stage voting was proposed to detect objects using template matching method while there are only a few samples. Firstly, the voting space was constructed off-line by using probability model through several samples. Then, a method of two-stage voting was used to detect objects in testing images. In the first stage, the components of object from testing image were detected, and the positions of components in query image were saved. In the second stage, the similarity of the object was computed integrally based on the components. According to the theory analysis and experimental results, the proposed method obtains lower computation complexity and higher precisions than previous works.

The relatively large time cost of the existing influence algorithms does not fit the social networks of which the scale keeps expanding. An influence optimization model was proposed based on the network community structure for solving problem of large time cost. Firstly evaluate nodes influence in each community and dig core members, and then find a small subset of nodes in the set composed of the core nodes and linking community nodes to get the maximization diffusion with minimization cost. The experimental results demonstrate that our model achieves the subset with more abroad influence diffusion and reduces running time compared with traditional methods. Its influence coverage is up to 90%.

Concerning the Maximum Transmission Unit (MTU) problems of nested mobile networks, based on the analysis of the existing mechanisms and the characteristics of the network structure, a tunnel MTU model and a tunnel MTU discovery mechanism based on location update were proposed. By storing the path MTU values between the home Agents at them, and adding the MTU information into the signaling messages, such as router advertisement and binding update, the mechanism can track the tunnel MTU fast and securely with the location update process, and adapt to the multihoming configuration and a variety of route optimization schemes. The simulation and analysis show that this mechanism can reduce packet delay and transmission overhead, and improve bandwidth utilization compared to the existing mechanisms.

Password-based authentication and key exchange protocol have been widely used in various network services due to easy memory of password. Unfortunately, password-based authentication scheme also suffers from attacks because of the low entropy of password. In the year 2011, Islam et al.（ISLAM SK H, BISWAS G P. Improved remote login scheme based on ECC. IEEE-International Conference on Recent Trends in Information Technology. Washington, DC: IEEE Computer Society, 2011： 1221-1226）proposed an improved remote login scheme based on Elliptic Curve Cryptography (ECC).Whereas, the scheme was vulnerable to stolen-verifier and impersonation attacks and failed to provide mutual authentication. Therefore, the authors proposed a password-based Remote Authentication with Key Agreement (RAKA) protocol using ECC to tackle the problems in Islam et al.'s scheme. RAKA was based on Elliptic Curve Discrete Logarithm Problem (ECDLP) and needed to compute six elliptic curve scale multiplications and seven hash function operations during a protocol run. The efficiency improves by about 15%〖BP(〗 percent〖BP)〗. It is more secure and efficient than Islam et al.'s scheme.

Session Initiation Protocol (SIP) provides authentication and session key agreement to ensure the security of the successive session. In 2010, Yoon et al. (YOON E-J, YOO K-Y. A three-factor authenticated key agreement scheme for SIP on elliptic curves. NSS '10: 4th International Conference on Network and System Security. Piscataway: IEEE, 2010: 334-339.) proposed a three-factor authenticated key agreement scheme named TAKASIP for SIP. However, the scheme is vulnerable to insider attack, server-spoofing attack, off-line password attack, and losing token attack. Moreover, it does not provide mutual authentication. To overcome these flaws of TAKASIP, a new three-factor authentication scheme named ETAKASIP based on Elliptic Curve Cryptosystem (ECC) was proposed. ETAKASIP, on the basis of elliptic curve discrete logarithm problem, provides higher security than TAKASIP. It needs 7 elliptic curve scalar multiplication operations, 1 additional operation and up to 6 Hash operations, and of high efficiency.

Concerning the collision and access failure which are very likely to occur when the intensity of the users’ access increases, a random access packetbased strategy was proposed. The users were divided into several groups according to the strategy; certain subframes formed a superframe to enable different users to send Uplink Synchronization Code (SYNCUL) to the corresponding subframe of superframe when the users needed random access information, and based on it, priority was set for the users from different groups. Compared with the general strategy, packetbased strategy has greatly improved the success rate of the users’ access, and grouping based on priority was adopted to ensure a higher success rate for advanced users. The theoretical analysis and simulation results show that the proposed packetbased strategy can significantly improve the system Quality of Service (QoS) and could be an effective measurement for decreasing the probability of collision in random access.

In order to manage the whole network, a novel traffic model was proposed based on the regression formula method. The core of model had three parts: (1) acquiring the delay by injecting probe packet train into the network based on Internet Protocol Measurement Protocol (IPMP); (2) deducing regression formula according to M/M/1 queue model; (3) acquiring the path by active measurement. In order to compute regression formula, the coefficients were acquired by the least square method. It is proved that the regression formula method is effective, and  the competed traffic error is within 30% by regress formula.