Pub Date : 2019-03-01DOI: 10.1109/CISS.2019.8692903
Daniel Salmond
Wireless communications networks can be modelled as cascades of emission events, which makes them amenable to being modelled as multivariate Hawkes processes (MHPs). The MHP parameters can then be used to evaluate an attributability matrix, which describes the probability that each emission event can be attributed to previous events. Methods for inferring probabilistic adjacency matrices and network throughput estimates from these attributability matrices are demonstrated.
{"title":"Blind estimation of wireless network topology and throughput","authors":"Daniel Salmond","doi":"10.1109/CISS.2019.8692903","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692903","url":null,"abstract":"Wireless communications networks can be modelled as cascades of emission events, which makes them amenable to being modelled as multivariate Hawkes processes (MHPs). The MHP parameters can then be used to evaluate an attributability matrix, which describes the probability that each emission event can be attributed to previous events. Methods for inferring probabilistic adjacency matrices and network throughput estimates from these attributability matrices are demonstrated.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122923268","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 : 2019-03-01DOI: 10.1109/CISS.2019.8692916
P. Jacquod, Laurent Pagnier
The energy transition’s ultimate goal is to meet energy demand from human activities sustainably. Accordingly, the penetration of new renewable energy sources (RES) such as photovoltaic panels and wind turbines is increasing in most power grids. In their current configuration, RES are essentially inertialess, therefore, low inertia situations are more and more common, in periods of high RES production, making grid stability a high concern in power grids with high share of RES. It has been suggested that the resulting reduction of overall inertia can be compensated to some extent by the deployment of substitution inertia-synthetic inertia, flywheels, synchronous condensers aso. Of particular importance is to optimize the placement of the limited available substitution inertia. Here, we construct a matrix perturbation theory to optimize inertia and primary control placement under the assumption that both are moderately heterogeneous. Armed with that efficient tool, we construct simple but efficient algorithms that independently determine the optimal geographical distribution of inertia and primary control.
{"title":"Optimal placement of inertia and primary control in high voltage power grids","authors":"P. Jacquod, Laurent Pagnier","doi":"10.1109/CISS.2019.8692916","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692916","url":null,"abstract":"The energy transition’s ultimate goal is to meet energy demand from human activities sustainably. Accordingly, the penetration of new renewable energy sources (RES) such as photovoltaic panels and wind turbines is increasing in most power grids. In their current configuration, RES are essentially inertialess, therefore, low inertia situations are more and more common, in periods of high RES production, making grid stability a high concern in power grids with high share of RES. It has been suggested that the resulting reduction of overall inertia can be compensated to some extent by the deployment of substitution inertia-synthetic inertia, flywheels, synchronous condensers aso. Of particular importance is to optimize the placement of the limited available substitution inertia. Here, we construct a matrix perturbation theory to optimize inertia and primary control placement under the assumption that both are moderately heterogeneous. Armed with that efficient tool, we construct simple but efficient algorithms that independently determine the optimal geographical distribution of inertia and primary control.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"167 ","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114057542","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 : 2019-03-01DOI: 10.1109/CISS.2019.8692863
Gaspar Tognetti, Jonah P. Sengupta, P. Pouliquen, A. Andreou
As computational needs increase in relation to the growing fields of Internet of Things and Deep Learning, energy-efficient, computational units are needed to bypass DSP units within Von Neumann architectures. A charge-mode vector matrix multiplier (VMM) with compute-in memory capabilities was fabricated in the Global Foundries 55nm LP process. The array is comprised of a 156 row by 512 column crossbar where each row computes a 512 element binary dot product in the charge domain. This normalized analog multiply and accumulate (MAC) is carried out by charge-injection devices who compute a 1-bit multiplication in the charge domain. Preliminary test results show successful, linear output computation in the analog domain to various input vectors that are both digital and multi-level analog. The 156 × 512 compute-in memory, CID array has been simulated to achieve an efficiency of 1.8 TOPs per mW.
{"title":"Characterization of a pseudo-DRAM Crossbar Computational Memory Array in 55nm CMOS : (Invited Paper)","authors":"Gaspar Tognetti, Jonah P. Sengupta, P. Pouliquen, A. Andreou","doi":"10.1109/CISS.2019.8692863","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692863","url":null,"abstract":"As computational needs increase in relation to the growing fields of Internet of Things and Deep Learning, energy-efficient, computational units are needed to bypass DSP units within Von Neumann architectures. A charge-mode vector matrix multiplier (VMM) with compute-in memory capabilities was fabricated in the Global Foundries 55nm LP process. The array is comprised of a 156 row by 512 column crossbar where each row computes a 512 element binary dot product in the charge domain. This normalized analog multiply and accumulate (MAC) is carried out by charge-injection devices who compute a 1-bit multiplication in the charge domain. Preliminary test results show successful, linear output computation in the analog domain to various input vectors that are both digital and multi-level analog. The 156 × 512 compute-in memory, CID array has been simulated to achieve an efficiency of 1.8 TOPs per mW.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114825014","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 : 2019-03-01DOI: 10.1109/CISS.2019.8692944
Liyuan Song, Qin Huang, Jiayi Rui
This paper proposes to analyze and construct minimum storage regenerating (MSR) codes for distributed storage systems based on their parity-check matrices. One codeword of constructed MSR codes is viewed as the superposition of several local codes and a global code, where the global code handles data reconstruction and the local codes deal with node recovery. Thus, the constructed MSR code can straightforward perform reconstruction-by-transfer. Moreover, by analyzing parity-check matrices of the local codes, it is sufficient to reveal the impact of code parameters of the MSR codes for single node failure, which coincides with the existing MSR codes.
{"title":"Construction and Analysis for Minimum Storage Regenerating Codes Based Parity-Check Matrices : Invited Presentation","authors":"Liyuan Song, Qin Huang, Jiayi Rui","doi":"10.1109/CISS.2019.8692944","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692944","url":null,"abstract":"This paper proposes to analyze and construct minimum storage regenerating (MSR) codes for distributed storage systems based on their parity-check matrices. One codeword of constructed MSR codes is viewed as the superposition of several local codes and a global code, where the global code handles data reconstruction and the local codes deal with node recovery. Thus, the constructed MSR code can straightforward perform reconstruction-by-transfer. Moreover, by analyzing parity-check matrices of the local codes, it is sufficient to reveal the impact of code parameters of the MSR codes for single node failure, which coincides with the existing MSR codes.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"422 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116087991","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 : 2019-03-01DOI: 10.1109/CISS.2019.8693058
M. E. Burich, R. Souza, J. Garcia-Frías
In this paper, we develop, for the first time in the literature, a Density Evolution analysis of Rate Compatible Modulation (RCM), which is challenging due to the way symbols in RCM are generated as weighted sums of the input bits. We consider uniform and non-uniform memoryless binary sources. By allowing the weights to be real numbers, rather than integers as in previous work, we propose, for the first time in the literature, an optimization procedure that automatically obtains the weights of the RCM scheme for a desired source entropy, signal to noise ratio, and information rate.
{"title":"Discretized Density Evolution for Rate Compatible Modulation : Invited Presentation","authors":"M. E. Burich, R. Souza, J. Garcia-Frías","doi":"10.1109/CISS.2019.8693058","DOIUrl":"https://doi.org/10.1109/CISS.2019.8693058","url":null,"abstract":"In this paper, we develop, for the first time in the literature, a Density Evolution analysis of Rate Compatible Modulation (RCM), which is challenging due to the way symbols in RCM are generated as weighted sums of the input bits. We consider uniform and non-uniform memoryless binary sources. By allowing the weights to be real numbers, rather than integers as in previous work, we propose, for the first time in the literature, an optimization procedure that automatically obtains the weights of the RCM scheme for a desired source entropy, signal to noise ratio, and information rate.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"137 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122850938","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 : 2019-03-01DOI: 10.1109/CISS.2019.8692838
S. Saab, Khaled Kamal Saab, S. Saab
Linear regression with shuffled labels is the problem of performing a linear regression fit on datasets whose labels are unknowingly shuffled with respect to their inputs. Such a problem relates to different applications such as genome sequence assembly, sampling and reconstruction of spatial fields, and communication networks. Existing methods are either applicable only to data with limited observation errors, work only for partially shuffled data, sensitive to initialization, and/or work only with small dimensions. This paper tackles this problem in its full generality using stochastic approximation, which is based on a first-order permutation-invariant constraint. We propose an optimal recursive algorithm that updates the estimate from the underdetermined function that is based on that permutation-invariant constraint. The proposed algorithm aims for per-iteration minimization of the mean square estimate error. Although our algorithm is sensitive to initialization errors, to the best of our knowledge, the resulting method is the first working solution for arbitrary large dimensions and arbitrary large observation errors while its computation throughput appears insignificant. Numerical simulations show that our method with shuffled datasets can outperform the ordinary least squares method without shuffling. We also consider a batch process to this problem where the datasets are independently available. The solution we propose is independent of initialization but requires that number of such datasets to be at least equal to the dimension of the unknown vector.
{"title":"Shuffled Linear Regression with Erroneous Observations","authors":"S. Saab, Khaled Kamal Saab, S. Saab","doi":"10.1109/CISS.2019.8692838","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692838","url":null,"abstract":"Linear regression with shuffled labels is the problem of performing a linear regression fit on datasets whose labels are unknowingly shuffled with respect to their inputs. Such a problem relates to different applications such as genome sequence assembly, sampling and reconstruction of spatial fields, and communication networks. Existing methods are either applicable only to data with limited observation errors, work only for partially shuffled data, sensitive to initialization, and/or work only with small dimensions. This paper tackles this problem in its full generality using stochastic approximation, which is based on a first-order permutation-invariant constraint. We propose an optimal recursive algorithm that updates the estimate from the underdetermined function that is based on that permutation-invariant constraint. The proposed algorithm aims for per-iteration minimization of the mean square estimate error. Although our algorithm is sensitive to initialization errors, to the best of our knowledge, the resulting method is the first working solution for arbitrary large dimensions and arbitrary large observation errors while its computation throughput appears insignificant. Numerical simulations show that our method with shuffled datasets can outperform the ordinary least squares method without shuffling. We also consider a batch process to this problem where the datasets are independently available. The solution we propose is independent of initialization but requires that number of such datasets to be at least equal to the dimension of the unknown vector.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129070016","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 : 2019-03-01DOI: 10.1109/CISS.2019.8693040
H. Nieto-Chaupis
We studied the possible formalisms used in both Quantum Electrodynamics and Classical Electrodynamics, that might be sharing same methodologies that turns out to be in quantization of the fields, despite of the fact that the field is an infinite wave. In one side we used the Volkov solutions while in the classical counterpart we used the formalism of Hartemann-Kerman. The obtained simulations would demonstrate that quantum and classical methodologies are based on the same mathematical basis that use the integer-order Bessel’s functions.
{"title":"Formalisms of Quantization in High Intensity Fields: Quantum Mechanics Meets Classical Electrodynamics","authors":"H. Nieto-Chaupis","doi":"10.1109/CISS.2019.8693040","DOIUrl":"https://doi.org/10.1109/CISS.2019.8693040","url":null,"abstract":"We studied the possible formalisms used in both Quantum Electrodynamics and Classical Electrodynamics, that might be sharing same methodologies that turns out to be in quantization of the fields, despite of the fact that the field is an infinite wave. In one side we used the Volkov solutions while in the classical counterpart we used the formalism of Hartemann-Kerman. The obtained simulations would demonstrate that quantum and classical methodologies are based on the same mathematical basis that use the integer-order Bessel’s functions.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"23 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128776356","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 : 2019-03-01DOI: 10.1109/CISS.2019.8692907
Keith Hermiston
Latin squares are combinatorial constructions that have found widespread application in communication systems through frequency hopping designs, error correcting codes and encryption algorithms. In this paper, a new, upper bound on the cardinality of the critical sets of all Latin squares of order n is presented. The bound is based on composite group structure and the summatory prime factorisation function (with multiplicities). The new bound aligns with all known, calculated cardinalities of largest critical sets. The proof addresses a long standing, open problem in discrete mathematics and impacts the assurance of systems based on Latin squares. The new bound also reveals a previously unknown, generative relationship between the smallest critical sets scs(n) and the largest critical sets lcs(n) of Latin squares.
{"title":"The Largest Critical Sets of Latin Squares","authors":"Keith Hermiston","doi":"10.1109/CISS.2019.8692907","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692907","url":null,"abstract":"Latin squares are combinatorial constructions that have found widespread application in communication systems through frequency hopping designs, error correcting codes and encryption algorithms. In this paper, a new, upper bound on the cardinality of the critical sets of all Latin squares of order n is presented. The bound is based on composite group structure and the summatory prime factorisation function (with multiplicities). The new bound aligns with all known, calculated cardinalities of largest critical sets. The proof addresses a long standing, open problem in discrete mathematics and impacts the assurance of systems based on Latin squares. The new bound also reveals a previously unknown, generative relationship between the smallest critical sets scs(n) and the largest critical sets lcs(n) of Latin squares.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"120 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124486775","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 : 2019-03-01DOI: 10.1109/CISS.2019.8692803
Li Li, M. Doroslovački, M. Loew
One consensus in the machine learning community is that obtaining good representations of the data is crucial for the classification tasks. But establishing a clear objective for representation learning is an open question and difficult. In this paper, we propose the Discriminant Analysis Loss Function (DALF) for the representation learning in Deep Neural Networks (DNNs). The gradients of DALF explicitly minimize the within-class variances (scatter) and maximize the between-class variances. We use DALF to drive the training of DNNs and call them Discriminant Analysis Deep Neural Networks (DisAnDNNs). Compared to other Linear Discriminant Analysis (LDA)-based cost functions, the computational cost of DALF is drastically reduced by avoiding eigen-decomposition and matrix inversion. We used simple datasets to illustrate the geometric meaning of DALF and compared it with LDA, then experimented with DALF-driven Residual Learning Nets (ResNets) on the pediatric pneumonia (chest X-ray image) dataset. The experimental results show that the DisAnDNNs achieve state-of the-art accuracy in the binary classification task. Particularly, in the pediatric pneumonia dataset, we achieved the accuracy of 96.63%, with a sensitivity of 99.23% and a specificity of 92.30%, all of which are better than the results in the literature that published the dataset.
{"title":"Discriminant Analysis Deep Neural Networks","authors":"Li Li, M. Doroslovački, M. Loew","doi":"10.1109/CISS.2019.8692803","DOIUrl":"https://doi.org/10.1109/CISS.2019.8692803","url":null,"abstract":"One consensus in the machine learning community is that obtaining good representations of the data is crucial for the classification tasks. But establishing a clear objective for representation learning is an open question and difficult. In this paper, we propose the Discriminant Analysis Loss Function (DALF) for the representation learning in Deep Neural Networks (DNNs). The gradients of DALF explicitly minimize the within-class variances (scatter) and maximize the between-class variances. We use DALF to drive the training of DNNs and call them Discriminant Analysis Deep Neural Networks (DisAnDNNs). Compared to other Linear Discriminant Analysis (LDA)-based cost functions, the computational cost of DALF is drastically reduced by avoiding eigen-decomposition and matrix inversion. We used simple datasets to illustrate the geometric meaning of DALF and compared it with LDA, then experimented with DALF-driven Residual Learning Nets (ResNets) on the pediatric pneumonia (chest X-ray image) dataset. The experimental results show that the DisAnDNNs achieve state-of the-art accuracy in the binary classification task. Particularly, in the pediatric pneumonia dataset, we achieved the accuracy of 96.63%, with a sensitivity of 99.23% and a specificity of 92.30%, all of which are better than the results in the literature that published the dataset.","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121124337","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 : 2019-03-01DOI: 10.1109/CISS.2019.8693025
Willie L. Thompson, Michael F. Talley
In recent years there has been a rapid evolution in research in the technology known as the Internet of Things or IoT. Consequently, this development has caused an increase in connected devices. According to Statista, the amount of IoT connected devices by the year 2025 will be 75.44 billion. Given this expected exponential rise in connected devices, this will cause an increase in the transmitted data by the year 2025 as well. The Data Management Solutions Review states that data creation will reach 163 zettabytes by 2025. These conditions will cause an escalation in data transmission which will cause problems such as latency, data rates, congestion, nonlinearities, and other complexities. While communication systems have performed well based on traditional mathematical transforms, there is a need to present new solutions to mitigate these problems. One potential solution is to resort to advanced Machine Learning (ML) techniques to help manage the rise in data volumes and algorithm-driven applications. The recent success of Deep Learning (DL) underpins new and powerful tools that tackle problems in this space. The unique parameters of DL techniques are capable of properly characterizing and categorizing complex signals being transmitted and received. This paper will investigate the optimization of communication systems at the physical layer (PHY) with future applications in IoT hardware implementation using a 1D Convolutional Neural Networks (CNN).
{"title":"Deep Learning for IoT Communications : Invited Presentation","authors":"Willie L. Thompson, Michael F. Talley","doi":"10.1109/CISS.2019.8693025","DOIUrl":"https://doi.org/10.1109/CISS.2019.8693025","url":null,"abstract":"In recent years there has been a rapid evolution in research in the technology known as the Internet of Things or IoT. Consequently, this development has caused an increase in connected devices. According to Statista, the amount of IoT connected devices by the year 2025 will be 75.44 billion. Given this expected exponential rise in connected devices, this will cause an increase in the transmitted data by the year 2025 as well. The Data Management Solutions Review states that data creation will reach 163 zettabytes by 2025. These conditions will cause an escalation in data transmission which will cause problems such as latency, data rates, congestion, nonlinearities, and other complexities. While communication systems have performed well based on traditional mathematical transforms, there is a need to present new solutions to mitigate these problems. One potential solution is to resort to advanced Machine Learning (ML) techniques to help manage the rise in data volumes and algorithm-driven applications. The recent success of Deep Learning (DL) underpins new and powerful tools that tackle problems in this space. The unique parameters of DL techniques are capable of properly characterizing and categorizing complex signals being transmitted and received. This paper will investigate the optimization of communication systems at the physical layer (PHY) with future applications in IoT hardware implementation using a 1D Convolutional Neural Networks (CNN).","PeriodicalId":123696,"journal":{"name":"2019 53rd Annual Conference on Information Sciences and Systems (CISS)","volume":"353 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-03-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134460482","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：