General Network Architecture Research Articles

DM-Fusion: Deep Model-Driven Network for Heterogeneous Image Fusion.

Heterogeneous image fusion (HIF) is an enhancement technique for highlighting the discriminative information and textural detail from heterogeneous source images. Although various deep neural network-based HIF methods have been proposed, the most widely used single data-driven manner of the convolutional neural network always fails to give a guaranteed theoretical architecture and optimal convergence for the HIF problem. In this article, a deep model-driven neural network is designed for this HIF problem, which adaptively integrates the merits of model-based techniques for interpretability and deep learning-based methods for generalizability. Unlike the general network architecture as a black box, the proposed objective function is tailored to several domain knowledge network modules to model the compact and explainable deep model-driven HIF network termed DM-fusion. The proposed deep model-driven neural network shows the feasibility and effectiveness of three parts, the specific HIF model, an iterative parameter learning scheme, and data-driven network architecture. Furthermore, the task-driven loss function strategy is proposed to achieve feature enhancement and preservation. Numerous experiments on four fusion tasks and downstream applications illustrate the advancement of DM-fusion compared with the state-of-the-art (SOTA) methods both in fusion quality and efficiency. The source code will be available soon.

Predicting the UV Escape Fraction of the First Galaxies in the Renaissance Simulations with Machine Learning

Cosmic reionization is likely driven by UV starlight emanating from the first generations of galaxies. A galaxy’s UV escape fraction, or the fraction of photons escaping from the galaxy, is useful to quantify its contribution to reionization. However, the UV escape fraction is notoriously difficult to predict due to local environment dependency and variability over time. Using data from the Renaissance Simulations, we attempt to make predictions about the impact of the first stars and galaxies on their environments. We present a time-independent classification model using a general artificial neural network architecture to predict the UV escape fraction given other galaxy properties—namely halo mass, stellar mass, redshift, star formation rate, lookback time, and gas fraction. We find our validation accuracy to be approximately 50%–65%, depending on the data set size from each zoom-in region of the Renaissance Simulations.

5G Wireless Networks

Everybody loves speed and moreover speedy internet, so its no surprise that every major telecom in the world is working to make it even faster. Smartphones, watches, homes, and cars are increasingly requiring stable internet connections. In order to survive in the world where in every second the speed changes and where we urge for more and more technology, here comes the fifth generation technology: 5G. In future, i.e., a world beyond 4G, some of the prime objectives that need to be fulfilled are increased capacity, improved data rate, decreased latency, and quality service. To meet these demands, large scale improvement in the cellular architecture of 5G is required .This paper basically lays emphasis on the 5th generation i.e. 5G cellular network architecture and some of the essential emerging technologies that can prove fruitful in humanizing the architecture and summiting the demands of users. This paper is contented with the details related to 5g with the prime focus on the massive multiple input multiple output technology and device-to-device communication (D2D). A general credible 5G cellular network architecture is being proposed with the guideline taken from the internet books and by the detailed study of the topic

Task-Specific Scene Structure Representations

Understanding the informative structures of scenes is essential for low-level vision tasks. Unfortunately, it is difficult to obtain a concrete visual definition of the informative structures because influences of visual features are task-specific. In this paper, we propose a single general neural network architecture for extracting task-specific structure guidance for scenes. To do this, we first analyze traditional spectral clustering methods, which computes a set of eigenvectors to model a segmented graph forming small compact structures on image domains. We then unfold the traditional graph-partitioning problem into a learnable network, named Scene Structure Guidance Network (SSGNet), to represent the task-specific informative structures. The SSGNet yields a set of coefficients of eigenvectors that produces explicit feature representations of image structures. In addition, our SSGNet is light-weight (56K parameters), and can be used as a plug-and-play module for off-the-shelf architectures. We optimize the SSGNet without any supervision by proposing two novel training losses that enforce task-specific scene structure generation during training. Our main contribution is to show that such a simple network can achieve state-of-the-art results for several low-level vision applications including joint upsampling and image denoising. We also demonstrate that our SSGNet generalizes well on unseen datasets, compared to existing methods which use structural embedding frameworks. Our source codes are available at https://github.com/jsshin98/SSGNet.

Future of 5G Wireless Systems: A Review

Everybody loves speed and moreover speedy internet, so it’s no surprise that every major telecom in the world is working to make it even faster. Smartphone’s, watches, homes, and cars are increasingly requiring stable internet connections. In order to survive in the world where in
 every second the speed changes and where we urge for more and more technology, here comes the fifth generation technology: 5G. In future, i.e., a world beyond 4G, some of the prime objectives that need to be fulfilled are increased capacity, improved data rate, decreased latency, and quality service. To meet these demands, large scale improvement in the cellular architecture of 5G is required .This paper basically lays emphasis on the 5th generation i.e. 5G cellular network architecture and some of the essential emerging technologies that can prove fruitful in humanizing the architecture and summiting the demands of users.
 This paper is contented with the details related to 5g with the prime focus on the massive multiple input multiple output technology and device-to-device communication (D2D). A general credible 5G cellular network architecture is being proposed with the guideline taken from the internet books and by the detailed study of the topic.
 Beginning with an exploration of the evolution from 4G to 5G, we dissect the technical intricacies that empower 5G networks, including spectrum allocation, massive MIMO, and network slicing.
 The paper navigates through the diverse spectrum of applications that 5G facilitates, from Enhanced Mobile Broadband (eMBB) .As 5G integrates with emerging technologies such as edge computing and artificial intelligence, we unravel the synergies that are reshaping industries and fostering innovation.

Uniformly convex neural networks and non-stationary iterated network Tikhonov (iNETT) method

We propose a non-stationary iterated network Tikhonov (iNETT) method for the solution of ill-posed inverse problems. The iNETT employs deep neural networks to build a data-driven regularizer, and it avoids the difficult task of estimating the optimal regularization parameter. To achieve the theoretical convergence of iNETT, we introduce uniformly convex neural networks to build the data-driven regularizer. Rigorous theories and detailed algorithms are proposed for the construction of convex and uniformly convex neural networks. In particular, given a general neural network architecture, we prescribe sufficient conditions to achieve a trained neural network which is component-wise convex or uniformly convex; moreover, we provide concrete examples of realizing convexity and uniform convexity in the modern U-net architecture. With the tools of convex and uniformly convex neural networks, the iNETT algorithm is developed and a rigorous convergence analysis is provided. Lastly, we show applications of the iNETT algorithm in 2D computerized tomography, where numerical examples illustrate the efficacy of the proposed algorithm.

Security in Wireless Communication Networks

Over the last decade, wireless networks have experienced significant growth due to the proliferation of mobile devices and new development of radio technologies. Since the very beginning, security for wireless communication networks has always been regarded as a critical issue. In recent research frenzy on 5G and V2X communications, security plays a pivotal role as wireless Internet access serves as critical infrastructures to facilitate the digital transcations. This book timely provides a comprehensive introduction to the security in wireless communication networks, with respect to not only general concepts, but also practical principles and cases in specific wireless communication systems. This book covers a wide range starting from the general wireless communication network architecture and basic concepts of network security to the mathematical background and practical techniques, and further to the specific security issues in different wireless networks, such as WLAN, LTE, 5G, etc. Comprehensively covering the security issues, security mechanisms, design principles, and techniques in wireless communication networks. The book intends to be a self-contained and one semester textbook on wireless security for either undergraduate senior level or graduate level courses. This book can also serve as a reference for practitioners and researchers in the field to get an idea of the current and future research focus in the emerging wireless security area.

Unveil the potential of siamese framework for visual tracking

Most of the existing Siamese tracking methods follow the overall framework of SiamRPN, adopting its general network architecture and the local and linear cross-correlation operation to integrate search and template features, which restricts the introduction of more sophisticated structures for expressive appearance representation as well as the further improvements on tracking performance. Motivated by the recent progresses in vision Transformer and MLP, we first explore to accomplish a global, nonlinear and scale-invariant similarity measuring manner called Dynamic Cross-Attention (DCA). Specifically, template features are first decomposed along the spatial and channel dimension and then the Transformer Encoders are applied to adaptively excavate the long-range feature interdependency, producing reinforced kernels. As the kernels are successively multiplied to the search feature map, similarity scores between all the pixels on feature maps are estimated at once while the spatial scale of search features remains constant. Furthermore, we redesign each part of our Siamese network to further remedy the framework limitation with the assistant of DCA. Comprehensive experimental results on large-scale benchmarks indicate that our Siamese method realizes the efficient feature extraction, aggregation, refinement and interaction, outperforming state-of-the-art trackers.

Obesity Prediction with EHR Data: A Deep Learning Approach with Interpretable Elements

Childhood obesity is a major public health challenge. Early prediction and identification of the children at an elevated risk of developing childhood obesity may help in engaging earlier and more effective interventions to prevent and manage obesity. Most existing predictive tools for childhood obesity primarily rely on traditional regression-type methods using only a few hand-picked features and without exploiting longitudinal patterns of children's data. Deep learning methods allow the use of high-dimensional longitudinal datasets. In this paper, we present a deep learning model designed for predicting future obesity patterns from generally available items on children's medical history. To do this, we use a large unaugmented electronic health records dataset from a large pediatric health system in the US. We adopt a general LSTM network architecture and train our proposed model using both static and dynamic EHR data. To add interpretability, we have additionally included an attention layer to calculate the attention scores for the timestamps and rank features of each timestamp. Our model is used to predict obesity for ages between 3-20 years using the data from 1-3 years in advance. We compare the performance of our LSTM model with a series of existing studies in the literature and show it outperforms their performance in most age ranges.

Evaluating Common Reconnaissance Tools and Techniques for Information Gathering

A reconnaissance attack is a commonly overlooked step in penetration testing but a critical step that could help increase the effectiveness of an attack on a target. However, it is a common attack faced by companies and institutions, among others. It enables the attacker or penetration tester to gain valuable information on the target and select the best tools and methods that would make the attack successful. This study aims to identify and review existing state-of-the-art methodology for reconnaissance attacks based on certain techniques and evaluation metrics which will be beneficial to professional, ethical hackers in selecting the best-fit tool for a successful reconnaissance attack and enlighten organizations and the general public of the potential harm of a successful reconnaissance attack. This study explored seven online databases, which include Springer, Elsevier, Wiley, IEEE, ACM, ArXiv and Google Scholar. A total of 1306 publications were retrieved. Several screening criteria were executed to select relevant articles. Finally, 19 articles were identified for in-depth analysis. A quantitative evaluation was conducted on the selected articles using two search strategies: Techniques and source. A Quantitative Analysis (QA) was conducted on the selected articles and the outcome based on existing reconnaissance tools shows that 95.2% of the tools allowed experts to gather information by running their necessary command from the command line. While 4.8% of the tools do not provide a command-line interface allowing users to launch it from the command line interface while specifying some parameters to customize how it runs. 61.9% of the tools are network-based as they can be used to gather about the target's network infrastructure such as protocols, ports, DNS, IP address, hosts and the general network architecture. At the same time, 28.5% could be classified as hybrid as they allow the attacker to gather system-based and network-based information. This study concludes with findings that show that some of the tools operate in a different capacity; the best-fit tool is massively dependent on the attacker or penetration tester and the scenario's situations. Therefore, a tool should be selected based on the user's preference and the attack style.

The Evolution of Quantum Key Distribution Networks: On the Road to the Qinternet

Quantum key distribution (QKD) constitutes a symmetric secret key negotiation protocol capable of maintaining information-theoretic security. Given the recent advances in QKD networks, they have evolved from academic research to some preliminary applications. A QKD network consists of two or more QKD nodes interconnected by optical fiber or free space links. The secret keys are negotiated between any pair of QKD nodes, and then they can be delivered to multiple users in various areas for ensuring long-term protection and forward secrecy. We commence by introducing the QKD basics, followed by reviewing the development of QKD networks and their implementation in practice. Subsequently, we describe the general QKD network architecture, its elements, as well as its interfaces and protocols. Next, we provide an in-depth overview of the associated physical layer and network layer solutions, followed by the standardization efforts as well as the application scenarios associated with QKD networks. Finally, we discuss the potential future research directions and provide design guidelines for QKD networks.

Inter-database validation of a deep learning approach for automatic sleep scoring.

Study objectivesDevelopment of inter-database generalizable sleep staging algorithms represents a challenge due to increased data variability across different datasets. Sharing data between different centers is also a problem due to potential restrictions due to patient privacy protection. In this work, we describe a new deep learning approach for automatic sleep staging, and address its generalization capabilities on a wide range of public sleep staging databases. We also examine the suitability of a novel approach that uses an ensemble of individual local models and evaluate its impact on the resulting inter-database generalization performance.MethodsA general deep learning network architecture for automatic sleep staging is presented. Different preprocessing and architectural variant options are tested. The resulting prediction capabilities are evaluated and compared on a heterogeneous collection of six public sleep staging datasets. Validation is carried out in the context of independent local and external dataset generalization scenarios.ResultsBest results were achieved using the CNN_LSTM_5 neural network variant. Average prediction capabilities on independent local testing sets achieved 0.80 kappa score. When individual local models predict data from external datasets, average kappa score decreases to 0.54. Using the proposed ensemble-based approach, average kappa performance on the external dataset prediction scenario increases to 0.62. To our knowledge this is the largest study by the number of datasets so far on validating the generalization capabilities of an automatic sleep staging algorithm using external databases.ConclusionsValidation results show good general performance of our method, as compared with the expected levels of human agreement, as well as to state-of-the-art automatic sleep staging methods. The proposed ensemble-based approach enables flexible and scalable design, allowing dynamic integration of local models into the final ensemble, preserving data locality, and increasing generalization capabilities of the resulting system at the same time.

Action detection using a neural network elucidates the genetics of mouse grooming behavior.

Automated detection of complex animal behaviors remains a challenging problem in neuroscience, particularly for behaviors that consist of disparate sequential motions. Grooming is a prototypical stereotyped behavior that is often used as an endophenotype in psychiatric genetics. Here, we used mouse grooming behavior as an example and developed a general purpose neural network architecture capable of dynamic action detection at human observer-level performance and operating across dozens of mouse strains with high visual diversity. We provide insights into the amount of human annotated training data that are needed to achieve such performance. We surveyed grooming behavior in the open field in 2457 mice across 62 strains, determined its heritable components, conducted GWAS to outline its genetic architecture, and performed PheWAS to link human psychiatric traits through shared underlying genetics. Our general machine learning solution that automatically classifies complex behaviors in large datasets will facilitate systematic studies of behavioral mechanisms.

Generative Adversarial Networks for Video-to-Video Domain Adaptation

Endoscopic videos from multicentres often have different imaging conditions, e.g., color and illumination, which make the models trained on one domain usually fail to generalize well to another. Domain adaptation is one of the potential solutions to address the problem. However, few of existing works focused on the translation of video-based data. In this work, we propose a novel generative adversarial network (GAN), namely VideoGAN, to transfer the video-based data across different domains. As the frames of a video may have similar content and imaging conditions, the proposed VideoGAN has an X-shape generator to preserve the intra-video consistency during translation. Furthermore, a loss function, namely color histogram loss, is proposed to tune the color distribution of each translated frame. Two colonoscopic datasets from different centres, i.e., CVC-Clinic and ETIS-Larib, are adopted to evaluate the performance of domain adaptation of our VideoGAN. Experimental results demonstrate that the adapted colonoscopic video generated by our VideoGAN can significantly boost the segmentation accuracy, i.e., an improvement of 5%, of colorectal polyps on multicentre datasets. As our VideoGAN is a general network architecture, we also evaluate its performance with the CamVid driving video dataset on the cloudy-to-sunny translation task. Comprehensive experiments show that the domain gap could be substantially narrowed down by our VideoGAN.

Cognitive Radio in Low Power Wide Area Network for IoT Applications: Recent Approaches, Benefits and Challenges

Some recent survey statistics suggest that low power wide area networks (LPWANs) are fast becoming the most prevalent communication platform used in many applications of the Internet of Things (IoT). However, because most LPWANs are generally deployed in the presently congested industrial, scientific, and medical bands, they are invariably plagued by problems associated with spectral congestion, such as increased interference, reduced data rates, and spectra inefficiency. These problems are solvable by integrating cognitive radio (CR) technologies in LPWAN (termed CR-LPWAN), for which some pioneering solutions now exist in the literature. Consequently, this article takes an early look at some of these pioneering efforts pertaining to the development of CR-LPWAN systems. We discuss a general network architecture and a physical layer front-end model suitable for CR-LPWAN systems. Then, some notable state-of-the-art approaches for CR-LPWAN systems are discussed. Potential advantages of CR-LPWAN systems for IoT-based applications are also presented, and this article closes with a few research challenges and future research directions in this regard. This article aims to serve as a starting point for most budding researchers who may be interested in the development of effective and efficient CR-LPWAN systems for the enhancement of different IoT-based applications.

High Performance Predictive Analytics in IoT

Internet of things (IoT) is a quick-moving gathering of web associated sensors implanted in a wide-extending assortment of physical articles. While things can be any physical item (energize or lifeless) on the planet, to which you could associate or implant a sensor. Sensors can take countless potential estimations. Sensors produce gigantic measures of new, organized, unstructured, ongoing information, and structures enormous information. IoT information is exceptionally huge and confused, which can give genuine-time setting and supposition data about genuine articles or nature. Among the different challenges that the present IoT is facing, the three prime areas of concern are, need of efficient framework to receive IoT data, a need of a new scalable parallel indexing technique for efficiently storing IoT data and securing IoT generated data at all the stages i.e. from the edge devices to the cloud. A new efficient framework is introduced, which can retrieve meaningful information from these IoT devices and efficiently index it. For processing such enormous real time data generated from IoT devices, new techniques are introducing which are scalable and secure. The research proposes a general IoT network architecture. It describes the interconnectivity among the different things such as sensors, receivers and cloud. The proposed architecture efficiently receives real time data from all the sensors. The prime focus is on the elimination of the existing issues in IoT. Along with this, the provision has to make for standard future proofing against these new proposed schemes.

Selective feature connection mechanism: Concatenating multi-layer CNN features with a feature selector

Different layers of deep convolutional neural networks(CNNs) can encode different-level information. High-layer features always contain more semantic information, and low-layer features contain more detail information. However, low-layer features suffer from the background clutter and semantic ambiguity. During visual recognition, the feature combination of the low-layer and high-level features plays an important role in context modulation. If directly combining the high-layer and low-layer features, the background clutter and semantic ambiguity may be caused due to the introduction of detailed information. In this paper, we propose a general network architecture to concatenate CNN features of different layers in a simple and effective way, called Selective Feature Connection Mechanism (SFCM). Low-level features are selectively linked to high-level features with a feature selector which is generated by high-level features. The proposed connection mechanism can effectively overcome the above-mentioned drawbacks. We demonstrate the effectiveness, superiority, and universal applicability of this method on multiple challenging computer vision tasks, including image classification, scene text detection, and image-to-image translation.

Towards general network architecture design criteria for negative gas adsorption transitions in ultraporous frameworks

Switchable metal-organic frameworks (MOFs) have been proposed for various energy-related storage and separation applications, but the mechanistic understanding of adsorption-induced switching transitions is still at an early stage. Here we report critical design criteria for negative gas adsorption (NGA), a counterintuitive feature of pressure amplifying materials, hitherto uniquely observed in a highly porous framework compound (DUT-49). These criteria are derived by analysing the physical effects of micromechanics, pore size, interpenetration, adsorption enthalpies, and the pore filling mechanism using advanced in situ X-ray and neutron diffraction, NMR spectroscopy, and calorimetric techniques parallelised to adsorption for a series of six isoreticular networks. Aided by computational modelling, we identify DUT-50 as a new pressure amplifying material featuring distinct NGA transitions upon methane and argon adsorption. In situ neutron diffraction analysis of the methane (CD4) adsorption sites at 111 K supported by grand canonical Monte Carlo simulations reveals a sudden population of the largest mesopore to be the critical filling step initiating structural contraction and NGA. In contrast, interpenetration leads to framework stiffening and specific pore volume reduction, both factors effectively suppressing NGA transitions.

Broadband Services Implementation by Using Survivable ATM Architecture

In Communication Networks, connectivity with Data Voice and Video Services are enhanced by means of NXN connectivity. These three domains play a vital role viz interconnection of users (Networking), Transition from one stage to another (Switching) and determining the path connectivity (Routing are main to enhance the single period to multiperiod network services. Survivability techniques focus on improving the network efficiency rather than redundancy. Hence this paper focuses on the General Fiber Network Architecture by using ATM (Asynchronous Transfer Mode) with SONET (Synchronous Optical Network and also enable its support enhancement by using uniform Multiplexing Switched Network (MSN) and Transfer Mode (TM). It refers to the networking, switching and routing by using multiplexing process in end-to- end communication. In the previous era many experiments that use ATM technology, by integrating digital transmission services in a network of fiber optics media (B-ISDN) was used. In this paper ATM standards, switches are used to conduct the experiments by using Optical Cross Connectivity (OCC), Optical Path Protection and Restoration (OPPR), Optical Network Integration Scheme (ONIS) are presented by using Survivable ATM Architecture (SATMA-BS).

Building an Otoscopic screening prototype tool using deep learning

BackgroundOtologic diseases are often difficult to diagnose accurately for primary care providers. Deep learning methods have been applied with great success in many areas of medicine, often outperforming well trained human observers. The aim of this work was to develop and evaluate an automatic software prototype to identify otologic abnormalities using a deep convolutional neural network.Material and methodsA database of 734 unique otoscopic images of various ear pathologies, including 63 cerumen impactions, 120 tympanostomy tubes, and 346 normal tympanic membranes were acquired. 80% of the images were used for the training of a convolutional neural network and the remaining 20% were used for algorithm validation. Image augmentation was employed on the training dataset to increase the number of training images. The general network architecture consisted of three convolutional layers plus batch normalization and dropout layers to avoid over fitting.ResultsThe validation based on 45 datasets not used for model training revealed that the proposed deep convolutional neural network is capable of identifying and differentiating between normal tympanic membranes, tympanostomy tubes, and cerumen impactions with an overall accuracy of 84.4%.ConclusionOur study shows that deep convolutional neural networks hold immense potential as a diagnostic adjunct for otologic disease management.