摘要:Recently,deep learning has been widely applied to video and image based identity recognition and authorization tasks,including face recognition and person identification.However,machine learning models,especially deep learning models,can be easily fooled by adversarial attacks,which may cause the identity recognition systems to make a wrong decision.Therefore,dependable identity recognition and authorization has become one of the hot topics currently.Recent advances on dependable identity recognition and authorization from both information space and physical space were presented,where the development of the attack models on face detection,face recognition,person re-identification,and face anti-spoofing as well as printable adversarial patches were introduced.The algorithms of visual identity anonymization and privacy protection were further discussed.Finally,the datasets,experimental protocols and performance of dependable identity recognition methods were summarized,and the possible directions in the future research were presented.
摘要:With the development of information and communication technology,large-scale data collection has vastly promoted the application of machine learning in various fields.However,the data involved in machine learning often contains a lot of personal private information,which makes privacy protection face new risks and challenges,and has attracted more and more attention.The current progress of the related laws,regulations and standards to the personal privacy protection and data safety in machine learning were summarized.The existing work on privacy protection for machine learning was presented in detail.Privacy protection algorithms usually have influence on the data quality,model performance and communication cost.Thus,the performance of the privacy protection algorithms should be comprehensively evaluated in multiple dimensions.The performance evaluation metrics for the privacy protection algorithms for machine learning were presented,given with the conclusion that the privacy preservation on machine learning needs to balance the data quality,model convergence rate and communication cost.
摘要:Facing the huge difference of business and various personalized experience requirements in the future,KPI (key performance indicator),KQI (key quality indicator),QoS,QoE and KBI (key business indicator) was discussed,as well as the possible optimal design methods,potential challenges and technical research directions based on KPI,KQI and KBI,which was a progressive and unified whole.Finally,it was pointed out that the current fairness index was a kind of “experience” guidance,and there was no specific feedback closed-loop mechanism to form a clear guidance for the design criteria.The traction from the perspective of KBI may be an optional final and normalized evaluation criterion.It was pointed out that to do these works well in depth and breadth,it needed interdisciplinary in-depth research,close collaboration and cross layer design between different nodes of the network.
关键词:mobile communication network;evaluation criterion;KPI;KQI;KBI
摘要:To alleviate the performance loss caused by pilot contamination in massive multiple-input multiple-output (MIMO) systems,a semi-dynamic pilot allocation scheme based on graph coloring was proposed.Firstly,a pilot reuse threshold was derived for users.Then,based on the pilot reuse threshold,users of each cell are divided into cell-center users and cell-edge users,where cell-center users were allowed to reuse the same pilot.Finally,based on the graph coloring theory,the pilot resources were allocated to cell-edge users with serious pilot interference by multi-cell coordination.The simulation results show that the proposed pilot allocation scheme reduces the pilot interference and improves the uplink achievable rate of the whole system effectively.
摘要:With the increasing number of social network users in recent years,text-based user sentiment analysis technology has been widely concerned and applied.However,data sparsity and low accuracy often reduce the accuracy and speed of emotion recognition methods.The user emotion Biterm topic model (US-BTM) was proposed which could find user preference and emotional tendency from the text of specific places,so as to effectively use Biterm for topic modeling.The strategy of user aggregation to form pseudo-documents was used,and word pairs were created for the whole corpus to solve the problems of data sparsity and short text.Then the topic was studied through the lexical co-occurrence model,so as to infer the topic with abundant corps-level information,and the purpose of accurately predicting the user’s interest,preference and emotion to the specific scene was achieved by analyzing the lexical matching set in the comment corpus under the specific scene and the emotion of the corresponding topic.The experimental results show that the method proposed can accurately capture users’ emotional tendency and correctly reveal users’ preference,which can be widely used in social network content description,recommendation,social network user interest description,semantic analysis and other fields.
摘要:Optical fiber eavesdropping is one of the major hidden dangers of power grid information security,but detection is difficult due to its high concealment.Aiming at the eavesdropping problems faced by communication networks,an optical fiber eavesdropping detection method based on machine learning was proposed.Firstly,seven-dimensions feature vector extraction method was designed based on the influence of eavesdropping on the physical layer of transmission.Then eavesdropping was simulated and experimental feature vectors were collected.Finally,two machine learning algorithms were used for classification detection and model optimization.Experiments show that the performance of the neural network classification is better than the K-nearest neighbor classification,and it can achieve 98.1% eavesdropping recognition rate in 10% splitting ratio eavesdropping.
摘要:Prolate spheroidal wave function (PSWF) signals are the best energy-gathering signals in the time-frequency domain,here is an exploratory application of PSWF signals to continuous phase modulation technology.Firstly,the process of generating CPM modulated signals based on PSWF signals was described.Secondly,a calculation method for the power spectral density of CPM modulated signals based on PSWF using the autocorrelation function method was given.Finally,combined with numerical calculation and analysis,the power spectral density and signal occupied bandwidth characteristics of minimum frequency shift keying,sinusoidal frequency shift keying,Gaussian filtered minimum shift keying and CPM modulation signal based on Gaussian function were compared.Numerical calculation results show that the PSWF signal can obtain CPM modulation signal with better spectral performance and energy concentration,compared with rectangular pulse,raised cosine pulse,Gaussian pulse and Gauss-like pulse as the baseband frequency modulation signal of CPM.
摘要:In a frequency division duplex communication system,precoding at the base station usually requires terminal feedback channel state information.In a distributed array system,the channel conditions between the terminal and different access nodes in the cell are different because multiple access nodes are arranged at different locations in the cell.When the resources for feedback channel state information are limited,the allocation method of feedback bit needs to be optimized to improve the overall performance of the system.In a multi-user distributed array system,an access node selection method based on distance threshold was adopted.Based on this method,the quantization characteristics of the random vector quantization codebook and the Taylor expansion method were used to derive the system quantization capacity loss,and then an approximate expression was given.Based on this expression,a feedback bit allocation method was proposed.Compared with other allocation methods,the method of this paper was more universal because it did not limit the number of access nodes selected by the user.The simulation results show that the strategy proposed is superior to the traditional equal-bit allocation scheme and can obtain better performance when the feedback resources are limited.
摘要:In order to improve the security in the process of information transmission,the current encryption methods were studied,and the research status was reported which increased the security of encryption by processing the key.According to the principle of segmenting and grouping data in the process of sending and encrypting,a method of multi-key encryption was designed:the segmenting key group and grouping key group were generated by the source key and random number,and the corresponding key was selected for each segmented and grouped data by its sequence number according to the matching algorithm.This method realized the encryption of multi-key in the process of information transmission.Finally,on the one hand,MATLAB was used to test and verify the encryption efficiency of this method.On the other hand,the security of this method was analyzed from the theoretical and mathematical perspective,comparing with the traditional method.It is concluded that this method can increase the security of the traditional method without too much cost much cost.
摘要:With the implementation of cryptography law,the application of cryptography in information system in China has been promoted to the legal level,which requires the application of public key infrastructure in the mutual trust of network entities to be strengthened.A scheme of constructing unified telecom public key infrastructure with strict hierarchical structure was proposed,that is,establishing a national telecom root CA as a trust anchor,and each major telecom operator becomes an independent sub-ca,forming a certificate trust chain of “national telecom root CA- telecom operator sub-ca” and providing PKI security services.This scheme can not only realize the mutual recognition between telecom operators,but also help unify the telecom public key infrastructure to become a globally trusted electronic authentication service provider,so as to have more influence and say in the formulation of international certification standards.
摘要:The technology and standardization of machine vision were focused on,and the attributes and relations of machine vision,computer vision,and human vision were analyzed.The machine vision related technologies and typical applications such as network information,security intelligent surveillance and intelligent transportation were introduced,as well as the latest development of international and domestic mainstream SDOs of machine vision technology,system and service.The development of the core machine vision technology,video coding for machine vision standards in VCM and DCM,and machine vision systems and application SDOs were elaborated.On this basis,the future development trend of machine vision technology and standardization was analyzed.
摘要:At present,the traditional Web security detection methods act on the input and output of the program,which can not prevent malicious code entering the program after being distorted and confused,and it is difficult to meet the new requirements of Web application security protection.Based on the in-depth analysis of the risk of traditional data flow monitoring methods,combined with the technical characteristics of rasp,a Web security detection method based on rasp was proposed.The rasp probe was embedded in the parameters of authority discrimination function,system command execution function and database operation function in Web application,and the change of data flow was detected in real-time at the code interpreter level.This method was implemented based on Java language.It was proved in the laboratory that this method is better than the traditional Web security detection method in accuracy and detection time.Finally,the deployment and application scenarios of this method were analyzed and proposed.
摘要:In view of the business scenario of bad information detection on the internet,the method of detection based on the text content of the website was discussed .Classical text analysis techniques were reviewed.The key technical features and two different usages of Bert model were introduced.The specific implementation scheme of using the feature extraction method to detect website bad information was described in detail,and was compared with the traditional TF-IDF model and word2vec+LSTM model.The validity of this method is verified.
摘要:As a powerful tool for operators to deal with means of new attack and ever-changing security threats,threat intelligence is also the core power to transform traditional passive defense into active defense.The importance of threat intelligence to operators was introduced,the productive process of threat intelligence was analyzed in detail,and finally the typical application scenarios of threat intelligence were given,which could provide technical reference for telecom operators in the practice of threat intelligence.
关键词:threat intelligence;active defense;productive process of threat intelligence;application scenarios of threat intelligence
摘要:In order to better promote the construction of China Telecom’s network and information security governance system,optimize the network and information security product service system,and promote the cultivation and development of network and information security core competence,domestic and foreign industry benchmark enterprises were selected to carry out practical case analysis.The advanced successful experience of benchmarking enterprises in the field of network and information security were summarized,and the current development status and future development trend in the field of network and information security were further combined,and targeted development suggestions and measures were put forward to improve the governance system of China’s telecommunications network and information security,improve the governance ability,as well as related business,and the development of products and services provides useful help.
关键词:industry benchmark;network and information security;practice analysis
摘要:In view of the current system architecture and technical solutions of 5G large-scale antennas,new technological developments and innovation directions were proposed,including low profile and light weight radiation element,high precision amplitude and phase calibration networks,all-in-one dense radiation arrays,and highly integrated antenna without reflection,antenna integrated filter,antenna and filter common module,mechanical phase shift and digital shaping combined,5G dual-band antenna,4G & 5G intelligent antenna,etc.For each technological innovation direction,an overview and useful description were given.
摘要:B2B business model in 5G era will become the main competition area for operators.Opportunities come along with challenges from investment,market and technology such as huge construction expenditure,fragmented market and difficulties of industry integration.Research on B2B operation cases of domestic and foreign operators and internet giants in the post 4G and 5G era had been done,from which “communication +” service model,industry coordinator model and platform model were investigated especially.Finally,three B2B operations models of domestic operators were proposed.
摘要:In the process of flattening metropolitan area network,in order to improve the reliability of MPLS L3VPN services,OLT devices can be planned as VRRP branch nodes.This design could probably induce traps of backhaul routes.Two separate methods of switching backhaul routes were proposed.One is a strict method based on checking the reachability of OLT devices,the other is a simplified method based on the priority of routes.The experimental results show that,handover latency is between 2~ 8 s.These proposed methods of switching backhaul routes,can effectively prevent traps of backhaul routes,and improve the reliability of services.
摘要:The traditional network operation and maintenance often focuses on the operation status of the network and equipment,which is difficult to solve the user's subjective perception.It is urgent to establish an end-to-end Internet application awareness management system.Focusing on the quality analysis of web browsing,the quality analysis framework and method of user perceived data based on big data technology was studied,a quality analysis system framework based on big data technology was designed and proposed,which was used to realize the quality analysis of Internet business.Taking an international website as an example,a diagnosis method of "two looks and five cuts" was proposed to delimit and trace the cause,which improves the efficiency of manual analysis and verifies the effectiveness of the proposed method.