Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570617109
Xi Zhang, Jingqing Wang, H. V. Poor
To support ultra-reliable low-latency communications (URLLC) for time-sensitive multimedia 5G wireless services, several advanced techniques, including statistical delay-bounded quality-of-service (QoS) provisioning and finite blocklength coding (FBC), have been developed to upper-bound both delay and error- rate. On the other hand, millimeter wave (mmWave) cell-free (CF) massive multi-input multi-output (m-MIMO) techniques, where a large number of distributed access points (APs) jointly serve all users at millimeter wave frequencies using the same time- frequency resources, has emerged as one of the key promising candidate techniques to significantly improve QoS performance in 5G networks. Leveraging the sparse scattering characteristics of mmWave wireless channels, the arrival traffic can be partitioned into parallel substreams using scattering-clusters based mmWave wireless channel model to reduce queuing delay. However, due to the complexity of analyzing queueing dynamics across clustered mmWave wireless channels for CF m-MIMO schemes, it is challenging to statistically guarantee QoS performance in terms of upper-bounding delay and error-rate. To overcome the above- mentioned problems, in this paper we propose a novel analytical model to quantitatively characterize stochastic QoS performance of delay and error-rate across clustered mmWave channels for CF m-MIMO schemes. In particular, we develop CF m-MIMO system models across clustered mmWave wireless channels. We also apply the Mellin transform to derive an upper bound on the delay violation probability using the spatial multiplexing queue model. Our simulation results validate and evaluate our proposed FBC based mmWave CF m-MIMO schemes under statistical delay/error-rate bounded QoS constraints.
{"title":"Statistical Delay/Error-Rate Bounded QoS Provisioning Across Clustered MmWave-Channels Over Cell-Free Massive MIMO Based 5G Mobile Wireless Networks in the Finite Blocklength Regime","authors":"Xi Zhang, Jingqing Wang, H. V. Poor","doi":"10.1109/CISS48834.2020.1570617109","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570617109","url":null,"abstract":"To support ultra-reliable low-latency communications (URLLC) for time-sensitive multimedia 5G wireless services, several advanced techniques, including statistical delay-bounded quality-of-service (QoS) provisioning and finite blocklength coding (FBC), have been developed to upper-bound both delay and error- rate. On the other hand, millimeter wave (mmWave) cell-free (CF) massive multi-input multi-output (m-MIMO) techniques, where a large number of distributed access points (APs) jointly serve all users at millimeter wave frequencies using the same time- frequency resources, has emerged as one of the key promising candidate techniques to significantly improve QoS performance in 5G networks. Leveraging the sparse scattering characteristics of mmWave wireless channels, the arrival traffic can be partitioned into parallel substreams using scattering-clusters based mmWave wireless channel model to reduce queuing delay. However, due to the complexity of analyzing queueing dynamics across clustered mmWave wireless channels for CF m-MIMO schemes, it is challenging to statistically guarantee QoS performance in terms of upper-bounding delay and error-rate. To overcome the above- mentioned problems, in this paper we propose a novel analytical model to quantitatively characterize stochastic QoS performance of delay and error-rate across clustered mmWave channels for CF m-MIMO schemes. In particular, we develop CF m-MIMO system models across clustered mmWave wireless channels. We also apply the Mellin transform to derive an upper bound on the delay violation probability using the spatial multiplexing queue model. Our simulation results validate and evaluate our proposed FBC based mmWave CF m-MIMO schemes under statistical delay/error-rate bounded QoS constraints.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"102 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116116233","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570610804
G. NaganandaK., Rick S. Blum, Alec Koppel
In this paper, we study the impact of the presence of byzantine sensors on the reduced-rank linear least squares (LS) estimator. A sensor network with N sensors makes observations of the physical phenomenon and transmits them to a fusion center which computes the LS estimate of the parameter of interest. It is well-known that rank reduction exploits the bias-variance tradeoff in the full-rank estimator by putting higher priority on highly informative content of the data. The low-rank LS estimator is constructed using this highly informative content, while the remaining data can be discarded without affecting the overall performance of the estimator. We consider the scenario where a fraction 0 < α < 1 of the N sensors are subject to data falsification attack from byzantine sensors, wherein an intruder injects a higher noise power (compared to the unattacked sensors) to the measurements of the attacked sensors.Our main contribution is an analytical characterization of the impact of data falsification attack of the above type on the performance of reduced-rank LS estimator. In particular, we show how optimally prioritizing the highly informative content of the data gets affected in the presence of attacks. A surprising result is that, under sensor attacks, when the elements of the data matrix are all positive the error performance of the low- rank estimator experiences a phenomenon wherein the estimate of the mean-squared error comprises negative components. A complex nonlinear programming-based recipe is known to exist that resolves this undesirable effect; however, the phenomenon is oftentimes considered very objectionable in the statistical literature. On the other hand, to our advantage this effect can serve to detect cyber attacks on sensor systems. Numerical results are presented to complement the theoretical findings of the paper.
{"title":"Reduced-rank Least Squares Parameter Estimation in the Presence of Byzantine Sensors","authors":"G. NaganandaK., Rick S. Blum, Alec Koppel","doi":"10.1109/CISS48834.2020.1570610804","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570610804","url":null,"abstract":"In this paper, we study the impact of the presence of byzantine sensors on the reduced-rank linear least squares (LS) estimator. A sensor network with N sensors makes observations of the physical phenomenon and transmits them to a fusion center which computes the LS estimate of the parameter of interest. It is well-known that rank reduction exploits the bias-variance tradeoff in the full-rank estimator by putting higher priority on highly informative content of the data. The low-rank LS estimator is constructed using this highly informative content, while the remaining data can be discarded without affecting the overall performance of the estimator. We consider the scenario where a fraction 0 < α < 1 of the N sensors are subject to data falsification attack from byzantine sensors, wherein an intruder injects a higher noise power (compared to the unattacked sensors) to the measurements of the attacked sensors.Our main contribution is an analytical characterization of the impact of data falsification attack of the above type on the performance of reduced-rank LS estimator. In particular, we show how optimally prioritizing the highly informative content of the data gets affected in the presence of attacks. A surprising result is that, under sensor attacks, when the elements of the data matrix are all positive the error performance of the low- rank estimator experiences a phenomenon wherein the estimate of the mean-squared error comprises negative components. A complex nonlinear programming-based recipe is known to exist that resolves this undesirable effect; however, the phenomenon is oftentimes considered very objectionable in the statistical literature. On the other hand, to our advantage this effect can serve to detect cyber attacks on sensor systems. Numerical results are presented to complement the theoretical findings of the paper.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123861928","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570627364
Zisheng Wang, Rick S. Blum
In recent years, there has been a surge in research and development efforts on vehicular ad-hoc networks (VANETs) with the objective to make driving safer. VANETS that share sensor data can provide tremendous improvements in this respect. Unfortunately, such VANETs are known for numerous security concerns and are vulnerable to cyber-attacks. In this paper we focus on studying cyber physical attacks on VANETs which share sensor data among vehicles to track important objects, an important emerging topic that has received little attention. We develop an appropriate VANET system model along with attack detection methods that can find any attack that impacts tracking of important objects like other vehicles or pedestrians regardless of how the attack is launched. This includes attacks modifying hardware, software, sensor data, communications or anything else. We have not seen any similar work. We illustrate these ideas with numerical results for a specific efficient distributed tracking algorithm. We describe a attack detection algorithm and numerically investigate the performance.
{"title":"Cybersecurity of Inference in Vehicular Ad-hoc Networks : Invited Presentation","authors":"Zisheng Wang, Rick S. Blum","doi":"10.1109/CISS48834.2020.1570627364","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570627364","url":null,"abstract":"In recent years, there has been a surge in research and development efforts on vehicular ad-hoc networks (VANETs) with the objective to make driving safer. VANETS that share sensor data can provide tremendous improvements in this respect. Unfortunately, such VANETs are known for numerous security concerns and are vulnerable to cyber-attacks. In this paper we focus on studying cyber physical attacks on VANETs which share sensor data among vehicles to track important objects, an important emerging topic that has received little attention. We develop an appropriate VANET system model along with attack detection methods that can find any attack that impacts tracking of important objects like other vehicles or pedestrians regardless of how the attack is launched. This includes attacks modifying hardware, software, sensor data, communications or anything else. We have not seen any similar work. We illustrate these ideas with numerical results for a specific efficient distributed tracking algorithm. We describe a attack detection algorithm and numerically investigate the performance.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128141491","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570617107
Jianqiao Chen, Xi Zhang
In this paper, we develop a novel sparse Bayesian learning (SBL) scheme for recovery of block-sparse uplink channels in time-division duplex (TDD) massive multi-input multi-output (MIMO) based 5G mobile wireless networks. We first introduce a pattern-coupled hierarchical Gaussian prior to characterize the sparse channel dependency among neighboring antennas, where the dependency coefficient is modeled by birth- death process. Then, to derive hyperparameters, which are employed to control the sparsity of channel coefficients, we exploit an expectation-maximization (EM) formulation to iteratively maximize a lower bound on the posterior probability. In the M-step, the particle swarm optimization (PSO) algorithm is employed to maximize the lower bound efficiently. Finally, we develop a belief propagation (BP)-based pattern-coupled sparse Bayesian learning (PC-SBL) algorithm, referred to as the BP-PC-SBL, to recover block-sparse uplink channels. Based on the factor graph of blocksparse uplink channels, we show that messages in BP satisfy complex Gaussian probability distribution. Therefore, we only need to update their means and variances when updating the messages. BP-PC-SBL algorithm provides precise approximations of matrix inversion as computed by the conventional SBL algorithm, which results in significantly improved computational efficiency. Our numerical analyses validate and evaluate the effectiveness of our proposed schemes and algorithms.
{"title":"Belief Propagation Pattern-Coupled Sparse Bayesian Learning for Non-Stationary Uplink Channel Estimation Over Massive-MIMO Based 5G Mobile Wireless Networks","authors":"Jianqiao Chen, Xi Zhang","doi":"10.1109/CISS48834.2020.1570617107","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570617107","url":null,"abstract":"In this paper, we develop a novel sparse Bayesian learning (SBL) scheme for recovery of block-sparse uplink channels in time-division duplex (TDD) massive multi-input multi-output (MIMO) based 5G mobile wireless networks. We first introduce a pattern-coupled hierarchical Gaussian prior to characterize the sparse channel dependency among neighboring antennas, where the dependency coefficient is modeled by birth- death process. Then, to derive hyperparameters, which are employed to control the sparsity of channel coefficients, we exploit an expectation-maximization (EM) formulation to iteratively maximize a lower bound on the posterior probability. In the M-step, the particle swarm optimization (PSO) algorithm is employed to maximize the lower bound efficiently. Finally, we develop a belief propagation (BP)-based pattern-coupled sparse Bayesian learning (PC-SBL) algorithm, referred to as the BP-PC-SBL, to recover block-sparse uplink channels. Based on the factor graph of blocksparse uplink channels, we show that messages in BP satisfy complex Gaussian probability distribution. Therefore, we only need to update their means and variances when updating the messages. BP-PC-SBL algorithm provides precise approximations of matrix inversion as computed by the conventional SBL algorithm, which results in significantly improved computational efficiency. Our numerical analyses validate and evaluate the effectiveness of our proposed schemes and algorithms.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"34 26","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"113933825","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570627767
Kush R. Varshney
Instilling trust in high-stakes applications of machine learning is becoming essential. Trust may be decomposed into four dimensions: basic accuracy, reliability, human interaction, and aligned purpose. The first two of these also constitute the properties of safe machine learning systems. The second dimension, reliability, is mainly concerned with being robust to epistemic uncertainty and model mismatch. It arises in the machine learning paradigms of distribution shift, data poisoning attacks, and algorithmic fairness. All of these problems can be abstractly modeled using the theory of mismatched hypothesis testing from statistical signal processing. By doing so, we can take advantage of performance characterizations in that literature to better understand the various machine learning issues.
{"title":"On Mismatched Detection and Safe, Trustworthy Machine Learning","authors":"Kush R. Varshney","doi":"10.1109/CISS48834.2020.1570627767","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570627767","url":null,"abstract":"Instilling trust in high-stakes applications of machine learning is becoming essential. Trust may be decomposed into four dimensions: basic accuracy, reliability, human interaction, and aligned purpose. The first two of these also constitute the properties of safe machine learning systems. The second dimension, reliability, is mainly concerned with being robust to epistemic uncertainty and model mismatch. It arises in the machine learning paradigms of distribution shift, data poisoning attacks, and algorithmic fairness. All of these problems can be abstractly modeled using the theory of mismatched hypothesis testing from statistical signal processing. By doing so, we can take advantage of performance characterizations in that literature to better understand the various machine learning issues.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"137 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132348748","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570614323
Weichang Wang, Lei Ying
This paper studies the problem of communications between aircraft and a control tower for aviation risk monitoring over wireless channels. The control tower needs to monitor the state of each aircraft in real time by receiving reports from the aircraft. Due to limited bandwidth, only a subset of aircraft can communicate with the control tower at the same time. This paper focuses on the problem of optimal scheduling of data transmissions to minimize the risk. We formulate the problem as learning states of parallel Markov chains where each Markov chain represents an aircraft, and the objective is to minimize the information entropy of all the aircraft. We propose an algorithm based on Whittle’s index and study the indexability of the problem for both single-state wireless channels and multi-state wireless channels. Our numerical evaluations show that our algorithm improves the accuracy of the estimations compared with the heuristic scheduling methods such as greedy and Round& Robin.
{"title":"Learning Parallel Markov Chains over Unreliable Wireless Channels","authors":"Weichang Wang, Lei Ying","doi":"10.1109/CISS48834.2020.1570614323","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570614323","url":null,"abstract":"This paper studies the problem of communications between aircraft and a control tower for aviation risk monitoring over wireless channels. The control tower needs to monitor the state of each aircraft in real time by receiving reports from the aircraft. Due to limited bandwidth, only a subset of aircraft can communicate with the control tower at the same time. This paper focuses on the problem of optimal scheduling of data transmissions to minimize the risk. We formulate the problem as learning states of parallel Markov chains where each Markov chain represents an aircraft, and the objective is to minimize the information entropy of all the aircraft. We propose an algorithm based on Whittle’s index and study the indexability of the problem for both single-state wireless channels and multi-state wireless channels. Our numerical evaluations show that our algorithm improves the accuracy of the estimations compared with the heuristic scheduling methods such as greedy and Round& Robin.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134052681","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570617362
Timur Luguev, Dominik Seuss, Jens-Uwe Garbas
Recent studies show that heart rate variability (HRV) is an important physiological characteristic that reflects physiological and affective states of a person. Advancements in the field of remote camera-based photoplethysmography has made possible measurement of cardiac signals using just the raw face videos. Most of existing studies of camera-based cardiovascular monitoring focus on just heart rate (HR) estimation, leaving more interesting case of remote HRV estimation out of scope. However, knowing only the average HR is not enough for affective sensing applications, and measurement of HRV is beneficial. We propose a new framework, which uses deep spatiotemporal networks for contactless HRV measurements from raw facial videos. The proposed framework employs data augmentation technique. It was evaluated on two multimodal databases that consists face videos with synchronized physiological signals. Experiments demonstrate the advantage of our deep learning based approach for HRV estimation. We also achieved promising results for inclusion remote HRV estimation in affective sensing applications.
{"title":"Deep Learning based Affective Sensing with Remote Photoplethysmography","authors":"Timur Luguev, Dominik Seuss, Jens-Uwe Garbas","doi":"10.1109/CISS48834.2020.1570617362","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570617362","url":null,"abstract":"Recent studies show that heart rate variability (HRV) is an important physiological characteristic that reflects physiological and affective states of a person. Advancements in the field of remote camera-based photoplethysmography has made possible measurement of cardiac signals using just the raw face videos. Most of existing studies of camera-based cardiovascular monitoring focus on just heart rate (HR) estimation, leaving more interesting case of remote HRV estimation out of scope. However, knowing only the average HR is not enough for affective sensing applications, and measurement of HRV is beneficial. We propose a new framework, which uses deep spatiotemporal networks for contactless HRV measurements from raw facial videos. The proposed framework employs data augmentation technique. It was evaluated on two multimodal databases that consists face videos with synchronized physiological signals. Experiments demonstrate the advantage of our deep learning based approach for HRV estimation. We also achieved promising results for inclusion remote HRV estimation in affective sensing applications.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"163 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114318886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/ciss48834.2020.9086271
{"title":"CISS 2020 TOC","authors":"","doi":"10.1109/ciss48834.2020.9086271","DOIUrl":"https://doi.org/10.1109/ciss48834.2020.9086271","url":null,"abstract":"","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114594957","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570627514
Eyal Nitzan, Topi Halme, H. Poor, V. Koivunen
Large-scale sensor networks are used in modern applications to perform statistical inference. In particular, multiple change-point detection using a sensor network is of interest in applications, such as Internet of Things and environmental monitoring. In this paper, we consider deterministic multiple change-point detection using a sensor network, in which each sensor observes a different data stream and communicates with a fusion center (FC). Due to communication limitations, the fusion center monitors only a subset of the sensors at each time slot. We propose a detection procedure that takes into account these limitations. In this procedure, the FC monitors the sensors with the highest cumulative sum values under the communication limitations. It is shown that the proposed procedure is scalable in the sense that it attains an average detection delay (ADD) that does not increase with the number of sensors, while controlling the false discovery rate. Using the proposed procedure, we identify and analyze the tradeoff between reducing the ADD and reducing the average number of observations drawn until the change-points are declared.
{"title":"Deterministic Multiple Change-Point Detection with Limited Communication","authors":"Eyal Nitzan, Topi Halme, H. Poor, V. Koivunen","doi":"10.1109/CISS48834.2020.1570627514","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570627514","url":null,"abstract":"Large-scale sensor networks are used in modern applications to perform statistical inference. In particular, multiple change-point detection using a sensor network is of interest in applications, such as Internet of Things and environmental monitoring. In this paper, we consider deterministic multiple change-point detection using a sensor network, in which each sensor observes a different data stream and communicates with a fusion center (FC). Due to communication limitations, the fusion center monitors only a subset of the sensors at each time slot. We propose a detection procedure that takes into account these limitations. In this procedure, the FC monitors the sensors with the highest cumulative sum values under the communication limitations. It is shown that the proposed procedure is scalable in the sense that it attains an average detection delay (ADD) that does not increase with the number of sensors, while controlling the false discovery rate. Using the proposed procedure, we identify and analyze the tradeoff between reducing the ADD and reducing the average number of observations drawn until the change-points are declared.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"29 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116871622","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-03-01DOI: 10.1109/CISS48834.2020.1570601577
Yuwei Sun, Ng S. T. Chong, H. Ochiai
With the rapid development of information technology, adaptation of an information system in industries and institutes has become more and more common. However, attacks like using zombie networks to access a host thus causing it to shut down are frequent in recent years. Domain names play a significant role in the connection with a server, considered as a key for detecting these attacks. In this paper, we propose a text-based method to convert domain names into numeric features, based on the term frequency and inverse document frequency (TF-IDF). Then we adopt the variational autoencoder (VAE) consisting of an encoder and a decoder, extracting hidden information from features. Moreover, through collapsing the Gaussian distribution of these features at the hidden layer to its mean, the distribution of domain names is visualized. After that, we adopt a supervised learning called Convolutional Neural Network (CNN) for the classification between the malicious and benign. We train the model using feature vectors from the VAE. At last, the scheme achieves a validation accuracy of 0.868 for the malicious domain names detection.
{"title":"Text-based Malicious Domain Names Detection Based on Variational Autoencoder And Supervised Learning","authors":"Yuwei Sun, Ng S. T. Chong, H. Ochiai","doi":"10.1109/CISS48834.2020.1570601577","DOIUrl":"https://doi.org/10.1109/CISS48834.2020.1570601577","url":null,"abstract":"With the rapid development of information technology, adaptation of an information system in industries and institutes has become more and more common. However, attacks like using zombie networks to access a host thus causing it to shut down are frequent in recent years. Domain names play a significant role in the connection with a server, considered as a key for detecting these attacks. In this paper, we propose a text-based method to convert domain names into numeric features, based on the term frequency and inverse document frequency (TF-IDF). Then we adopt the variational autoencoder (VAE) consisting of an encoder and a decoder, extracting hidden information from features. Moreover, through collapsing the Gaussian distribution of these features at the hidden layer to its mean, the distribution of domain names is visualized. After that, we adopt a supervised learning called Convolutional Neural Network (CNN) for the classification between the malicious and benign. We train the model using feature vectors from the VAE. At last, the scheme achieves a validation accuracy of 0.868 for the malicious domain names detection.","PeriodicalId":256370,"journal":{"name":"2020 54th Annual Conference on Information Sciences and Systems (CISS)","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117100791","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：