Architecture Setup Research Articles

BERT with an Augmented Cross-attention Decoder (BERT-ACD) for Binary and Fine-grained Multiband Sentiment Detection

In contemporary times, research in sentiment analysis has taken deeper steps into a finer and more granular analysis, transcending beyond the traditional binary or ternary classification of sentiment/opinion into positive, negative, or neutral. With the increasing complexity and challenging nature of such tasks, large language models inspired by transformer architecture are frequently deployed to address such challenges. Despite recorded improvements, challenges in identifying different levels, strengths or bands of sentiment intensity and the aspect for which the sentiment is expressed remain unresolved. In this article, we propose a banded sentiment analysis system for categorizing texts into 7 meaningful and relatable bands of sentiment for modern applications. It is also capable of performing aspect-based sentiment analysis in the same pipeline. The system architecture is inspired by the transformer language model with a BERT-based encoder and a newly proposed cross-attention, non-autoregressive decoder with augmented inputs. The decoder receives an n-gram-based augmented input sequence embedding that is specifically extracted from the original input, which comprises a list of the subjects, descriptive phrases, and modification phrases that underscore cases of amplification or undertone in a sentence. Rule-based tree parsing was proposed for use with dependency parsing for the extraction of these augmented inputs for the cross-attention decoder. Extensive experiments were conducted under different architecture setups and conditions with popular sentiment analysis datasets (Amazon reviews 2023, IMDB Movies review, SST-5 and SST-2 datasets) to verify the efficacy of the system. Extended labeling was also performed on the SST-5 dataset to generate 7 sentiment classes with the help of GPT4 and Bard. Experiments validate the efficacy of the proposed models.

Lightweight MobileNet Model for Image Tempering Detection

In recent years, there has been a wide range of image manipulation identification challenges and an overview of image tampering detection and the relevance of applying deep learning models such as CNN and MobileNet for this purpose. The discussion then delves into the construction and setup of these models, which includes a block diagram as well as mathematical calculations for each layer. A literature study on Image tampering detection is also included in the discussion, comparing and contrasting various articles and their methodologies. The study then moves on to training and assessment datasets, such as the CASIA v2 dataset, and performance indicators like as accuracy and loss. Lastly, the performance characteristics of the MobileNet and CNN designs are compared. This work focuses on Image tampering detection using convolutional neural networks (CNNs) and the MobileNet architecture. We reviewed the MobileNet architecture's setup and block diagram, as well as its application to Image tampering detection. We also looked at significant literature on Image manipulation detection, such as major studies and their methodologies. Using the CASIA v2 dataset, we evaluated the performance of MobileNet and CNN architectures in terms of accuracy and loss. This paper offered an overview of the usage of deep learning and CNN architectures for image tampering detection and proved their accuracy in detecting manipulated images.

Cloud service for authentication of a person based on their electrocardiogram

Telemedicine grows faster with each year. In scope of it new technologies have been created to solve information communication problems. One of them is distance monitoring which requires electrocardiogram analysis. In this case it is important to confirm that the transferred electrocardiogram through information channels belongs to the patient. Researchers from different countries work on this problem. They suggest different methods of authentication by electrocardiogram. The goal of this work is to suggest a prototype of a service that could be used to authenticate electrocardiograms. The paper describes which algorithms have been used to build authentication technology and how it was implemented. There is a short history of built applications. It shows their structures and the purpose. The recently developed system is a prototype to authentication service. It performs registration of new electrocardiograms which is the most time consuming process in authentication. The paper describes the architecture of the system and shows the result of executed experiments. The results show that there is a performance issue with the machine learning library ML.NET. When a lot of cores are allocated to one machine learning instance the overheads highly decrease the overall experiment time. These experiments confirmed Amdahl’s law. Nevertheless, an architecture was found where experiments took the least time for execution. Knowing the issue with the machine learning library, a new architecture setup was suggested and will be implemented in future works. Besides, attention is paid on how developed service should help researchers to improve the technology. The main idea of the system is using one service for developing and testing the technology. That means we can be sure that the cloud service that runs on production would behave the same as in the research phase.

On the Relative Impact of Optimizers on Convolutional Neural Networks with Varying Depth and Width for Image Classification

The continued increase in computing resources is one key factor that is allowing deep learning researchers to scale, design and train new and complex convolutional neural network (CNN) architectures in terms of varying width, depth, or both width and depth to improve performance for a variety of problems. The contributions of this study include an uncovering of how different optimization algorithms impact CNN architectural setups with variations in width, depth, and both width/depth. Specifically in this study, three different CNN architectural setups in combination with nine different optimization algorithms—namely SGD vanilla, with momentum, and with Nesterov momentum, RMSProp, ADAM, ADAGrad, ADADelta, ADAMax, and NADAM—are trained and evaluated using three publicly available benchmark image classification datasets. Through extensive experimentation, we analyze the output predictions of the different optimizers with the CNN architectures using accuracy, convergence speed, and loss function as performance metrics. Findings based on the overall results obtained across the three image classification datasets show that ADAM and NADAM achieved superior performances with wider and deeper/wider setups, respectively, while ADADelta was the worst performer, especially with the deeper CNN architectural setup.

Trans-kingdom interactions in mixed biofilm communities.

The microbial world represents a phenomenal diversity of microorganisms from different kingdoms of life, which occupy an impressive set of ecological niches. Most, if not all, microorganisms once colonize a surface develop architecturally complex surface-adhered communities, which we refer to as biofilms. They are embedded in polymeric structural scaffolds and serve as a dynamic milieu for intercellularcommunication through physical and chemical signalling. Deciphering microbial ecology of biofilms in various natural or engineered settings has revealed coexistence of microorganisms from all domains of life, including Bacteria, Archaea, and Eukarya. The coexistence of these dynamic microbes is not arbitrary, as a highly coordinated architectural setup and physiological complexity show ecological interdependence and myriads of underlying interactions. In this review, we describe how species from different kingdoms interact in biofilms and discuss the functional consequences of such interactions. We highlight metabolic advances of collaboration among species from different kingdoms, and advocate that these interactions are of great importance and need to be addressed in future research. Since trans-kingdom biofilms impact diverse contexts, ranging from complicated infections to efficient growth of plants, future knowledge within this field will be beneficial for medical microbiology, biotechnology, and our general understanding of microbial life in nature.

Meaning-making strategies on the balcony: The night sky in K. Taha’s Beschreibung einer Krabbenwanderung (‘Portrait of a wandering crab’)

The meaning-making system of the sky has persisted for millennia (Peters 2015). Yet, many of us do not relate to the importance of the connection with celestial sky anymore. One of the many contributing factors is our relation to urban systems, as well as densification and, consequently, the verticalization of living. As an example of the power of place-making strategies and its friction with the architectural set-up in high-rise mass dwellings and estates, the novel Beschreibung einer Krabbenwanderung by Karosh Taha makes use of the night-time sky in idiosyncratic spaces of the high-rise, such as the balcony. Here, the links between relations of time and space, the here and there (), as well as inside and outside spaces, become apparent and show a connection of the liminality of post-migrant identities. In an analysis of the access to the night-time sky and its concluding place-making strategies by one of the protagonists of the novel, I show how the balcony functions as a liminal space that is idiosyncratic for the mobile and mediatized identity in the western European high-rise dwelling.

IIoT platforms’ architectural features – a taxonomy and five prevalent archetypes

In the industrial Internet of Things (IIoT), digital platforms have recently received significant attention. Although IIoT platforms revolve around similar business objectives, they address various use cases and, thus, differ considerably in their architectural setup. While research has already investigated IIoT platforms from a business or design perspective, little is known about their underlying technology stack and its implications. To unveil different IIoT platform configurations and better understand their architectural design, we systematically develop and validate a taxonomy of IIoT platforms’ architectural features based on related literature, real-world cases, and expert interviews. On this foundation, we identify and discuss five IIoT platform archetypes (Allrounder, Device Controller, Data Hub, Service Enabler, Connector). Our findings contribute to the descriptive knowledge in this ambiguous research field while also elucidating the interplay of IIoT platforms’ architectural setup and their purpose. From a managerial viewpoint, our results may guide practitioners in comparing and selecting a suitable IIoT platform.

RV-TEE: secure cryptographic protocol execution based on runtime verification

Analytical security of cryptographic protocols does not immediately translate to operational security due to incorrect implementation and attacks targeting the execution environment. Code verification and hardware-based trusted execution solutions exist, however these leave it up to the implementer to assemble the complete solution, imposing a complete re-think of the hardware platforms and software development process. We rather aim for a comprehensive solution for secure cryptographic protocol execution, which takes the form of a trusted execution environment based on runtime verification and stock hardware security modules. RV-TEE can be deployed on existing platforms and protocol implementations. Runtime verification lends itself well at several conceptual levels of the execution environment, ranging from high level protocol properties, to lower level checks such as taint inference. The proposed architectural setup involving two runtime verification modules is instantiated through a case study using a popular web browser. We successfully monitor high and low level properties with promising results with respect to practicality.

Blockchain-based Secure Data Transmission for UAV Swarm using Modified Particle Swarm Optimization Path Planning Algorithm

With the rapid development of unmanned aerial vehicles assisted applications enabled with a communication system, they are open to various malicious attacks. As a new form of flying things, they can access the network for better communication via the aerial base station. Most of the Unmanned aerial vehicles assisted flying objects' optimal path selection schemes does not consider the path deviation. In path deviation attacks, secure data transmission are not addressed in existing works. The secure communication process between Unmanned aerial vehicles and base station are exploited through security-based attacks. Moreover, path loss issue leads to multicast packet loss and unsecured broadcast. The existing network architecture setup does not fulfill the secure data communication and privacy issues. In this paper, Blockchain is utilized to investigate the secure communication between Unmanned aerial vehicles to Wireless Unmanned aerial vehicles Base stations. Since the destination information is dynamic under an uncertain environment, it will cause a delay in data communication. Unmanned aerial vehicles are more vulnerable to security attacks. The proposed blockchain-based architecture supports secure data communication in Unmanned aerial vehicles uncertain environments. To improve network security, this paper designs a modified particle swarm optimization method for better path selection. Through these experimental results, a blockchain-based data communication scheme is outer performed concerning network security.

Exploring Mission-Oriented Innovation Ecosystems for Sustainability: Towards a Literature-Based Typology

With mounting sustainability challenges, policy makers have embraced the idea of transformative, mission-oriented innovation policies, to direct innovation objectives towards the ‘grand challenges’ in recent years. Against this backdrop, the discourse on innovation ecosystems, bringing together actors from science, industry, government and civil society for collaborative research and innovation, has increasingly gained traction. Yet, their role and architectural set-up in a sustainability context remains rather poorly understood. Complementing a systematic literature review with methods of bibliometric analysis and typology building, this paper introduces a typology of mission-oriented innovation ecosystems. It finds that, depending on the type of mission they are trying to address, ecosystems differ, with both a view to the actors involved, and the specific role taken on by them throughout the innovation process. In particular, it points to an increasingly important role of the state for realizing system-level transformations, underlines the importance of civil society involvement, and highlights research organizations’ need to adapt to new requirements.

Shallow or Deep? Detecting Anomalous Flows in the Canadian Automated Clearing and Settlement System using an Autoencoder

Financial market infrastructures and their participants play a crucial role in the economy. Financial or operational challenges faced by one participant can have contagion effects and pose risks to the broader financial system. Our paper applies (deep) neural networks (autoencoder) to detect anomalous flows from payments data in the Canadian Automated Clearing and Settlement System (ACSS) similar to Triepels et al. (2018). We evaluate several neural network architecture setups based on the size and number of hidden layers, as well as differing activation functions dependent on how the input data was normalized. As the Canadian financial system has not faced bank runs in recent memory, we train the models on “normal” data and evaluate out-of-sample using test data based on historical anomalies as well as simulated bank runs. Our out-of-sample simulations demonstrate the autoencoder’s performance in different scenarios, and results suggest that the autoencoder detects anomalous payment flows reasonably well. Our work highlights the challenges and trade-offs in employing a workhorse deep-learning model in an operational context and raises policy questions around how such outlier signals can be used by the system operator in complying with the prominent payment systems guidelines and by financial stability experts in assessing the impact on the financial system of a financial institution that shows extreme behaviour.

Arquitecturas de la industrialización nacional. El conjunto Papeles Bío-Bío como recinto autónomo primario (1953 hasta la actualidad)

In the mid-20th century –which marked the beginning of the industrialization process in Chile– several firms critical for development established themselves on the banks of the Biobio River, in what would later become a strategic location. Many of these companies connected the factory facilities with housing complexes for their workers. This work takes as case study the Bio-Bio Papers complex, located in the current commune of San Pedro de la Paz, Greater Concepcion, Chile. Its objective is to highlight the architectural-urban morphology of the case’s industrial complex. A historical study is conducted which collects information from secondary sources to then focus on the description and analysis of its urban and architectural setup, further delving into its spatial values.

Towards Energy-Efficient Heterogeneous Multicore Architectures for Edge Computing

In recent years, the edge computing paradigm has been attracting much attention in the Internet-of-Things domain. It aims to push the frontier of computing applications, data, and services away from the usually centralized cloud servers to the boundary of the network. The benefits of this paradigm shift include better reactivity and reliability, reduced data transfer costs toward the centralized cloud servers, and enhanced confidentiality. The design of energy-efficient edge compute nodes requires, among others, low power cores such as microprocessors. Heterogeneous architectures are key solutions to address the crucial energy-efficiency demand in modern systems. They combine various processors providing attractive power and performance trade-offs. Unfortunately, no standard heterogeneous microcontroller-based architecture exists for edge computing. This paper deals with the aforementioned issue by exploring typical low power architectures for edge computing. Various heterogeneous multicore designs are developed and prototyped on FPGA for unbiased evaluation. These designs rely on cost-effective and inherently ultra-low power cores commercialized by Cortus SA, a world-leading semiconductor IP company in the embedded ultra-low power microcontroller domain. Some microarchitecture-level design considerations, e.g., floating point and out-of-order computing capabilities, are taken into account for exploring candidate solutions. In addition, a tailored and flexible multi-task programming model is defined for the proposed architecture paradigm. We analyze the behavior of various application programs on available core configurations. This provides valuable insights on the best architecture setups that match program characteristics, so as to enable increased energy-efficiency. Our experiments on multi-benchmark programs show that on average 22% energy gain can be achieved (up to 45%) compared to a reference system design, i.e., a system with the same execution architecture, but agnostic of the task management insights gained from the comprehensive evaluation carried out in this work.

Complex Predicates and Multidimensionality in Grammar

This paper contributes to the on-going discussion of how best to analyze and handle complex predicate formations, commenting in particular on the properties of Hindi N-V complex predicates as set out by Vaidya et al. (2019). I highlight features of existing LFG analyses and focus in particular on the modular architecture of LFG, its attendant multidimensional lexicon and the analytic consequences which follow from this. I point out where the previously existing LFG proposals have been misunderstood as viewed from the lens of theories such as LTAG and HPSG, which assume a very different architectural set-up and provide a comparative discussion of the issues.

Hadoop‐based analytic framework for cyber forensics

SummaryWith an exponential increase in the data size and complexity of various documents to be investigated, existing methods of network forensics are found not much efficient with respect to accuracy and detection ratio. The existing techniques for network forensic analysis exhibit inherent limitations while processing a huge volume, variety, and velocity of data. It makes network forensic a time‐consuming and resource‐consuming task. To balance time taken and output delivered, these existing techniques put a limit on the amount of data under analysis, which results in a polynomial time complexity of these solutions. So to mitigate these issues, in this paper, we propose an effective framework to overcome the limitation to handle large volume, variety, and velocity of data. An architectural setup that consists of MapReduce framework on top of Hadoop Distributed File System environment is proposed in this paper. The proposed framework demonstrates its capability to handle issues of storage and processing of big data using cloud computing. Also, in the proposed framework, supervised machine learning (random forest‐based decision tree) algorithm has been implemented to demonstrate better sensitivity. To train and validate the model, online available data set from CAIDA is taken and university network traffic samples, with increasing size, has been taken for experiment. Results thus obtained confirm the superiority of the proposed framework in network forensics, with an average accuracy of 99.34% (malicious and nonmalicious traffic).

An Efficient Industrial Big-Data Engine

Current trends in industrial systems opt for the use of different big-data engines as a means to process huge amounts of data that cannot be processed with an ordinary infrastructure. The number of issues an industrial infrastructure has to face is large and includes challenges such as the definition of different efficient architecture setups for different applications, and the definition of specific models for industrial analytics. In this context, this paper explores the development of a medium size big-data engine (i.e., implementation) able to improve the performance in map-reduce computing by splitting the analytic into different segments that may be processed by the engine in parallel using a hierarchical model. This type of facility reduces end-to-end computation time for all segments with their results then merged with other information from other segments after their processing in parallel. This type of setup increases the performance of current clusters improving I/O operations remarkably as empirical results revealed.

Off-the-Shelf and Flexible Hybrid Frequency and Time Domain Experimental Architecture Setup for Electrochemical Energy Modules Testing Under Realistic Operating Conditions

This paper proposes a standard and flexible test bench architecture for the testing and modeling of electrochemical energy modules. It also presents a comprehensive methodology that enables to obtain accurate models under realistic operating conditions with a low computational cost. For this purpose, three different configuration modes of the experimental architecture setup are introduced. Finally, the test bench architecture and the experimental methodology proposed have been validated by the modeling and testing of a Li-ion battery pack and a supercapacitor module.

Scalable Anti-Censorship Framework Using Moving Target Defense for Web Servers

Although the Internet has become a hub around which every aspect of our lives—from commerce to leisurely activities—is centered, many around the world are not able to freely access information over the Internet. Some governments censor what the people can and cannot see. In this paper, regardless of the socio-political view points, we focus on the design of anti-censorship technology that can be implemented on the side of the information purveyors. The primary objective is to develop a framework for combating censorship. Our approach aims to make it too expensive and impractical for the adversary to censor Web sites. In particular, we propose the use of Mobile IPv6 to form a moving target defense strategy, where the Web servers logically behave as if they are the mobile nodes (without actually moving). The potential efficacy of this framework is modeled analytically. Probabilistic models are used to derive important metrics and parameters. One key factor termed swarming ratio enables hosting sites to reason about the amount of resources needed to force the adversary’s costs over practical limits. This model is used to guide the performance goals and architectural setup of the prototype implementation (modifications are made on the server-side software and Kernel without changing the standard Mobile IPv6 protocol). Hence, the solution can be utilized without any changes to the existing network infrastructure. Furthermore, we introduce a novel, credit-based accounting strategy for grouping of users to drastically shift resource requirements in our favor. Lab-based tests are used to measure performance overheads, and based on the findings, targeted optimizations are performed to consider practical deployment scenarios. The end result is a solution that may also be combined with existing anti-censorship methods (that are end-user-based and/or assisted by friendly network assets) to form a robust anti-censorship solution.

Ultra-Low Power Dynamic Knob in Adaptive Compressed Sensing Towards Biosignal Dynamics.

Compressed sensing (CS) is an emerging sampling paradigm in data acquisition. Its integrated analog-to-information structure can perform simultaneous data sensing and compression with low-complexity hardware. To date, most of the existing CS implementations have a fixed architectural setup, which lacks flexibility and adaptivity for efficient dynamic data sensing. In this paper, we propose a dynamic knob (DK) design to effectively reconfigure the CS architecture by recognizing the biosignals. Specifically, the dynamic knob design is a template-based structure that comprises a supervised learning module and a look-up table module. We model the DK performance in a closed analytic form and optimize the design via a dynamic programming formulation. We present the design on a 130 nm process, with a 0.058 mm (2) fingerprint and a 187.88 nJ/event energy-consumption. Furthermore, we benchmark the design performance using a publicly available dataset. Given the energy constraint in wireless sensing, the adaptive CS architecture can consistently improve the signal reconstruction quality by more than 70%, compared with the traditional CS. The experimental results indicate that the ultra-low power dynamic knob can provide an effective adaptivity and improve the signal quality in compressed sensing towards biosignal dynamics.

A novel rhamno-mannan exopolysaccharide isolated from biofilms of Burkholderia multivorans C1576

Burkholderia multivorans C1576 is a Gram negative opportunistic pathogen causing serious lung infection in cystic fibrosis patients. Considering that bacteria naturally form biofilms, and exopolysaccharides are recognized as important factors for biofilm architecture set-up, B. multivorans was grown both in biofilm and in non-biofilm mode on two different media in order to compare the exopolysaccharides biosynthesized in these different experimental conditions. The exopolysaccharides produced were purified and their structure was determined resorting mainly to NMR spectroscopy, ESI mass spectrometry and gas chromatography coupled to mass spectrometry. The experimental data showed that both in biofilm and non-biofilm mode B. multivorans C1576 produced a novel exopolysaccharide having the following structure: [Display omitted] About 50% of the 2-linked rhamnose residues are substituted on C-3 with a methyl ether group.The high percentage of deoxysugar Rha units, coupled with OMe substitutions, suggest a possible role for polymer domains with marked hydrophobic characteristics able to create exopolysaccharide junction zones favouring the stability of the biofilm matrix.