Unit Architecture Research Articles

Selective-Hop Graph Neural Networks for semi-supervised node classification

Message-passing Graph Neural Networks (GNNs) are remarkably successful in graph representation learning. Most of them propagate and aggregate information within fixed-size receptive-fields (RFs) for all nodes. In this paper, our empirical studies discover that different-sized RFs are often desired for distinct graphs and target nodes. To this end, we propose a dynamic selection mechanism by allowing each node to adaptively adjust the size of its RFs according to the graph properties. In particular, we design an architectural unit called Selective-Hop (SH) block, in which information from neighborhoods within different hops are merged using dual attention mechanisms: (i) node-wise attention on each network branch to capture mutual information between the subgraphs and the entire graph; and (ii) channel-wise attention on different network branches to detect the significance of different hops at the node-level granularity. As such, the SH block determines neighbors flexibly without the constraints of hops. By stacking several SH blocks in GNN style, we instantiate two new GNN variants, SHGCN and SHSGC. Empirical results on eight benchmark datasets for semi-supervised node classification reveal that our new models achieve a considerable improvement over the state-of-the-art methods under low-homophily or heterophily and maintain competitive performance under homophily.

Single-clan settlements and multiple-clan settlements: a case study based on the evolution of spatial fractals in four traditional villages in the Xianggan region of China

ABSTRACT Compared with those in other Chinese provinces, the villages in Hunan and Jiangxi are characterized by complex clan networks. Our purpose was to elucidate the complexity and hierarchical changes in the evolution of clan-organized settlement patterns and the spatial fractal structure of traditional villages. Box-counting dimension analysis, which is based on fractal theory, was used to evaluate the evolution of four villages in Hunan and Jiangxi, China. Our findings indicated that (1) both single-clan and multiclan villages exhibit a spatial fractal pattern; (2) the fractal dimension of individual buildings is greater than that of building groups and the village as a whole; and (3) the self-similarity fractal dimension is greater for single-clan villages than for multiclan villages. These findings establish a connection between the restoration needs of traditional villages and the potential application of the self-similarity principle. They support a hierarchical approach to conservation, beginning with individual architectural units to preserve the villages’ original authenticity. Moreover, the observed fractal differentiation revealed the fundamental clan-cultural logic underlying traditional village spaces. This logic elucidates the potential of leveraging each village’s unique clan culture to counteract the homogenization of village space development, thereby revitalizing the grassroots dynamics of village spatial development.

Multiple and Gyro-Free Inertial Datasets

An inertial navigation system (INS) utilizes three orthogonal accelerometers and gyroscopes to determine platform position, velocity, and orientation. There are countless applications for INS, including robotics, autonomous platforms, and the internet of things. Recent research explores the integration of data-driven methods with INS, highlighting significant innovations, improving accuracy and efficiency. Despite the growing interest in this field and the availability of INS datasets, no datasets are available for gyro-free INS (GFINS) and multiple inertial measurement unit (MIMU) architectures. To fill this gap and to stimulate further research in this field, we designed and recorded GFINS and MIMU datasets using 54 inertial sensors grouped in nine inertial measurement units. These sensors can be used to define and evaluate different types of MIMU and GFINS architectures. The inertial sensors were arranged in three different sensor configurations and mounted on a mobile robot, a passenger car and a turntable. In total, the dataset contains 45 hours of inertial data and corresponding ground truth trajectories. The data is freely accessible through our figshare repository.

Architecture designing of digital twin in a healthcare unit.

Objectives: This study proposes a novel architecture for designing digital twins in healthcare units. Methods: A systematic research methodology was employed to develop architecture design patterns. In particular, a systematic literature review was conducted using the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) framework to answer specific research questions and provide guidelines for designing the architecture. Subsequently, a case study was designed and analyzed at a chemotherapy treatment center for outpatients. Results: System architecture knowledge was distilled from this real-world case study, supplemented by existing software and systems design patterns. A novel five-layer architecture for digital twins in healthcare units was proposed with a focus on the security and privacy of patients' information. Conclusion: The proposed digital twin architecture for healthcare units offers a comprehensive solution that provides modularity, scalability, security, and interoperability. The architecture provides a robust framework for effectively and efficiently managing healthcare environments.

Recognition of Real-Time Video Activities Using Stacked Bi-GRU with Fusion-based Deep Architecture

Recognizing and understanding human activities in real-time videos is a challenging task due to the complex nature of video data and the need for efficient and accurate analysis. This research pioneers a breakthrough in video activity recognition by introducing a robust framework leveraging the power of a stacked Bidirectional Long Short-Term Memory (Bi-LSTM) and Gated Recurrent Unit (GRU) architecture, harmonized within a fusion-based deep model. The stacked Bi-LSTM-GRU model capitalizes on its dual recurrent architecture, capturing nuanced temporal dependencies within video sequences. The fusion-based deep architecture synergizes spatial and temporal features, enabling the model to discern intricate patterns in human activities. To further enhance the discriminative power of the model, we introduce a fusion module in the proposed deep architecture. The fusion module integrates multi-modal features extracted from different levels of the network hierarchy, allowing for a more comprehensive representation of video activities. We demonstrate the efficacy of our approach through rigorous experimentation on UCF50, UCF101, and HMDB51 datasets. In experiments on the UCF50 dataset, our model achieves an accuracy of 97.01% and 95.86% on training and validation sets respectively, showcasing its proficiency in discerning activities across a diverse range of scenarios. The evaluation extends to the UCF101 dataset, where the proposed approach achieves a competitive accuracy of 97.62% and 96.93% on training and validation sets, surpassing previous benchmarks by a margin of approx 1%. Further-more, on the challenging HMDB51 dataset, the model demonstrates a robust accuracy of 89.71%and 88.88% on training and validation sets, solidifying its efficacy in intricate action recognition tasks.

TOURISM MANAGEMENT EXPLAINED BY THE CORRELATION BETWEEN THE AVERAGE TOTAL INCOME PER PERSON AND THE NUMBER OF OVERNIGHT STAYS IN ACCOMMODATION UNITS

In Romania’s tourism industry, in the last decade, following the European trend, small dimensioned accommodation units have appeared, bearing the name of boutique guesthouses or boutique hotels, depending on their accommodation capacity. The boutique guesthouses offer both accommodation services in rooms with traditional Romanian boyar characteristics, with Saxon and Szechuan influences, as well as food services with the same culinary characteristics. A particularity of these accommodation units is that each room had been furnishing according to a traditional noble theme, using either furniture that has been produced in Romanian workshops. Roomꞌ s names have been inspired by the woody essences that can be found in the Romanian forests and a large part of the furniture is made of cherry, beech, fir, oak, elm wood. The specific architecture of the accommodation units has been established in 18th century specific to the rococo and baroque style, predominantly found in buildings constructions in Romanian towns where niche tourism is a small part of mass tourism industry.

Basin-scale exhumed Cretaceous fluvial systems (Chubut Group, Cañadón Asfalto basin), central Patagonia, Argentina

This contribution describes basin-scale, Cretaceous (Late Aptian to Cenomanian-Early Turonian) fluvial systems outcropped over ∼30,000 km2 that contain more than 7000 exhumed paleochannels in the Chubut Group, Cañadón Asfalto basin (central Patagonia), Argentina. Paleochannels are encased within a ∼350 m thick stratigraphic interval (∼20 My) including volcaniclastic and epiclastic fluvial units (Cerro Barcino and Puesto Manuel Arce formations). Geomorphologically, the exhumed paleochannels conform to isolated sandstones/conglomeratic ridges, a swarm of paleochannels, and elongated/equidimensional mesas (plateaus), characterizing much of the “mesetiform” landscape in central Chubut. The plan-view analysis of exhumed paleochannels mainly shows a mean WNW-ESE orientation (112°–292°) and predominance of narrow (83.1%) and very narrow (14.4%), low-sinuosity channels (96%). A plan-view multi-scale architectural analysis of sandbodies allows us to recognize different fluvial hierarchies: a) 3rd-order lithosomes (macroform growth increments), b) 4th-order lithosomes (point bars and crevasse-channels), b) 5th-order lithosomes (main channels), and c) 6th-order lithosomes (channel belts). Two main types of channel arrangements in the exhumed fluvial systems were recognized: fixed-channel and mobile channel belts. We observed a broad relationship between geomorphologic configurations and fluvial architectural units. Some macro spatiotemporal variation in the fluvial architecture was observed towards the youngest positions, including wider channels and a greater relative abundance of high-sinuosity channels with point bars. The basin-scale distribution of exhumed paleochannels would allow us to infer the predominance of shallow lacustrine or lagoon deposits in the western basin region, probably controlled by the inherited topography of the Jurassic- Early Cretaceous depocenters.

Enhancing image caption generation through context-aware attention mechanism

Image captioning, the process of generating natural language descriptions based on image content, has garnered attention in AI research for its implications in scene understanding and human-computer interaction. While much prior research has focused on caption generation for English, addressing low-resource languages like Bengali presents challenges, particularly in producing coherent captions linking visual objects with corresponding words. This paper proposes a context-aware attention mechanism over semantic attention to accurately diagnose objects for image captioning in Bengali. The proposed architecture consists of an encoder and a decoder block. We chose ResNet-50 over the other pre-trained models for encoding the image features due to its ability to solve the vanishing gradient problem and recognize complex object features. For decoding generated captions, a bidirectional Gated Recurrent Unit (GRU) architecture combined with an attention mechanism captures contextual dependencies in both directions, resulting in more accurate captions. The paper also highlights the challenge of transferring knowledge between domains, especially with culturally specific images. Evaluation of three Bengali benchmark datasets, namely BAN-Cap, BanglaLekhaImageCaption, and Bornon, demonstrates significant performance improvement in METEOR score over existing methods by approximately 30%, 18%, and 45%, respectively. The proposed context-aware, attention-based image captioning system significantly outperforms current state-of-the-art models in Bengali caption generation despite limitations in reference captions on certain datasets.

Semi-Supervised Left-Atrial Segmentation Based on Squeeze–Excitation and Triple Consistency Training

Convolutional neural networks (CNNs) have achieved remarkable success in fully supervised medical image segmentation tasks. However, the acquisition of large quantities of homogeneous labeled data is challenging, making semi-supervised training methods that rely on a small amount of labeled data and pseudo-labels increasingly popular in recent years. Most existing semi-supervised learning methods, however, underestimate the importance of the unlabeled regions during training. This paper posits that these regions may contain crucial information for minimizing the model’s uncertainty prediction. To enhance the segmentation performance of the left-atrium database, this paper proposes a triple consistency segmentation network based on the squeeze-and-excitation mechanism (SETC-Net). Specifically, the paper constructs a symmetric architectural unit called SEConv, which adaptively recalibrates the feature responses in the channel direction by modeling the inter-channel correlations. This allows the network to adaptively weigh each channel according to the task’s needs, thereby emphasizing or suppressing different feature channels. Moreover, SETC-Net is composed of an encoder and three slightly different decoders, which convert the prediction discrepancies among the three decoders into unsupervised loss through a constructed iterative pseudo-labeling scheme, thus encouraging consistent and low-entropy predictions. This allows the model to gradually capture generalized features from these challenging unmarked regions. We evaluated the proposed SETC-Net on the public left-atrium (LA) database. The proposed method achieved an excellent Dice score of 91.14% using only 20% of the labeled data. The experiments demonstrate that the proposed SETC-Net outperforms seven current semi-supervised methods in left-atrium segmentation and is one of the best semi-supervised segmentation methods on the LA database.

Ultra-Wideband Phased Antenna Array Time Delay Unit Architecture Optimization in Presence of Component Non-Idealities

Ultra-Wideband Phased Antenna Array Time Delay Unit Architecture Optimization in Presence of Component Non-Idealities

A novel phase-field lattice Boltzmann framework for diffusion-driven multiphase evaporation

Heat transfer and phase change phenomena, particularly diffusion-driven droplet evaporation, play pivotal roles in various industrial applications and natural processes. Despite advancements in computational fluid dynamics, modeling multiphase flows with large density ratios remains challenging. In this study, we developed a robust and stable conservative Allen–Cahn-based phase-field lattice Boltzmann method to solve the flow field equations. This method is coupled with the finite difference discretization of vapor species transport equation and the energy equation. The coupling between the vapor concentration and temperature field at the interface is modeled by the well-known Clausius–Clapeyron correlation. Our approach is capable of simulations under real physical conditions and is compatible with graphics processing unit architecture, making it ideal for large-scale industrial simulations. Three validation test cases are conducted to demonstrate the consistency of the presented model, including simulations of Stefan flow, the evaporation of suspended droplets containing water, acetone, and ethanol in the air, and the evaporation of a water sessile droplet on a flat surface. The results show that the model is able to predict the behavior and characteristics of each case accurately. Notably, our numerical results exhibit a maximum relative error of approximately 1% in simulations of Stefan flow. In the case of suspended droplet evaporation, the observed maximum difference between the calculated wet bulb temperatures and those derived from psychrometric charts is approximately 0.9 K. Moreover, our analysis of the sessile droplet reveals a good agreement between the results obtained by our model for the evaporative mass flux and those obtained from the existing models in the literature for different contact angles.

A United Architecture For AES/Present Ciphers and Its Usage in a Soc Environment

Abstract: In order to monitor the operating status of vehicles, it is necessary to collect vehicle operating data in real time through IoT devices and analyze these data. However, the collected data has the characteristics of multi-source heterogeneous, network resources are limited and server performance is poor. It is difficult to truly realize data processing in real time. In addition, data needs to be transmitted over the network, it is particularly important to ensure the safety of data transmission. Considering the above problems, it is necessary to structure the data and define a unified data format to facilitate data transmission and analysis. At the same time, improve the server communication program and improve the server's concurrent processing capabilities. In addition, considering that data needs to be transmitted over the network, in order to ensure that the data is not stolen or tampered with, the PRESENT lightweight encryption algorithm is adopted to ensure the safety of data transmission. Compared with encryption algorithms such as AES, this algorithm has much lower hardware requirements. This article combines the characteristics of the project and uses the number of communications between the device and the server to achieve the dynamic key update which is approximately one-time pad, and greatly improves the security of the data.

Compensation adaptive robust control for a linear motor–driven stage system with state and input constraints based on gated recurrent unit architecture

Compensation adaptive robust control for a linear motor–driven stage system with state and input constraints based on gated recurrent unit architecture

An Interpretable Light Attention–Convolution–Gate Recurrent Unit Architecture for the Highly Accurate Modeling of Actual Chemical Dynamic Processes

To equip data-driven dynamic chemical process models with strong interpretability, we develop a light attention–convolution–gate recurrent unit (LACG) architecture with three sub-modules—a basic module, a brand-new light attention module, and a residue module—that are specially designed to learn the general dynamic behavior, transient disturbances, and other input factors of chemical processes, respectively. Combined with a hyperparameter optimization framework, Optuna, the effectiveness of the proposed LACG is tested by distributed control system data-driven modeling experiments on the discharge flowrate of an actual deethanization process. The LACG model provides significant advantages in prediction accuracy and model generalization compared with other models, including the feedforward neural network, convolution neural network, long short-term memory (LSTM), and attention-LSTM. Moreover, compared with the simulation results of a deethanization model built using Aspen Plus Dynamics V12.1, the LACG parameters are demonstrated to be interpretable, and more details on the variable interactions can be observed from the model parameters in comparison with the traditional interpretable model attention-LSTM. This contribution enriches interpretable machine learning knowledge and provides a reliable method with high accuracy for actual chemical process modeling, paving a route to intelligent manufacturing.

Compact and Low-Latency FPGA-Based Number Theoretic Transform Architecture for CRYSTALS Kyber Postquantum Cryptography Scheme

In the modern era of the Internet of Things (IoT), especially with the rapid development of quantum computers, the implementation of postquantum cryptography algorithms in numerous terminals allows them to defend against potential future quantum attack threats. Lattice-based cryptography can withstand quantum computing attacks, making it a viable substitute for the currently prevalent classical public-key cryptography technique. However, the algorithm’s significant time complexity places a substantial computational burden on the already resource-limited chip in the IoT terminal. In lattice-based cryptography algorithms, the polynomial multiplication on the finite field is well known as the most time-consuming process. Therefore, investigations into efficient methods for calculating polynomial multiplication are essential for adopting these quantum-resistant lattice-based algorithms on a low-profile IoT terminal. Number theoretic transform (NTT), a variant of fast Fourier transform (FFT), is a technique widely employed to accelerate polynomial multiplication on the finite field to achieve a subquadratic time complexity. This study presents an efficient FPGA-based implementation of number theoretic transform for the CRYSTAL Kyber, a lattice-based public-key cryptography algorithm. Our hybrid design, which supports both forward and inverse NTT, is able run at high frequencies up to 417 MHz on a low-profile Artix7-XC7A100T and achieve a low latency of 1.10μs while achieving state-of-the-art hardware efficiency, consuming only 541-LUTs, 680 FFs, and four 18 Kb BRAMs. This is made possible thanks to the newly proposed multilevel pipeline butterfly unit architecture in combination with employing an effective coefficient accessing pattern.

Efficient protein structure archiving using ProteStAr.

The introduction of Deep Minds' Alpha Fold 2 enabled the prediction of protein structures at an unprecedented scale. AlphaFold Protein Structure Database and ESM Metagenomic Atlas contain hundreds of millions of structures stored in CIF and/or PDB formats. When compressed with a general-purpose utility like gzip, this translates to tens of terabytes of data, which hinders the effective use of predicted structures in large-scale analyses. Here, we present ProteStAr, a compressor dedicated to CIF/PDB, as well as supplementary PAE files. Its main contribution is a novel approach to predicting atom coordinates on the basis of the previously analyzed atoms. This allows efficient encoding of the coordinates, the largest component of the protein structure files. The compression is lossless by default, though the lossy mode with a controlled maximum error of coordinates reconstruction is also present. Compared to the competing packages, i.e. BinaryCIF, Foldcomp, PDC, our approach offers a superior compression ratio at established reconstruction accuracy. By the efficient use of threads at both compression and decompression stages, the algorithm takes advantage of the multicore architecture of current central processing units and operates with speeds of about 1 GB/s. The presence of Python and C++ API further increases the usability of the presented method. The source code of ProteStAr is available at https://github.com/refresh-bio/protestar.

Applying seismic sedimentology to characterize the architecture of a fan delta complex: Triassic Baikouquan formation, Mahu Sag, Junggar Basin, western China

Reservoir architecture is often heterogeneous and can affect the migration and accumulation of hydrocarbons. The Triassic Baikouquan Formation in the Mahu oilfield presents a unique opportunity to use seismic data for fan delta channel and lobe belt characterization in an area with a sparse well spacing of about 800–1200 m. The Baikouquan Formation comprises coarse-grained lacustrine fan delta units deposited by traction and debris flows. The occurrence, distribution, and morphology were investigated using petrography, frequency spectrum decomposition, stratal slices, seismic forward modeling, and wireline logs. Fan-shaped lobe complexes and distributary channel complexes (sixth-order architectural units) exhibiting relatively large lateral and vertical variation were identified from 45- and 60-Hz seismic profiles. Channel and fan-shaped lobe belts (fifth-order architectural units) were investigated by integrating provenances, architectural elements from individual wells, the ratio of conglomerate to formation, cross section analysis, and seismic stratal slices. Four channel belts and three fan-shaped lobe belts were recognized, occurring in six evolutionary stages. It was discovered that these channel and lobe belts had distinct geomorphological traits. Reservoir architecture and distribution was governed by paleogeomorphology, fluctuations in accommodation space, and sediment supply. The identified fifth-order architectural units are useful for reservoir modeling and contribute to remaining oil development. This study indicates the suitability of seismic sedimentology for characterizing reservoir architecture in areas with sparse well spacing.

New proposal for the Architectural Stratigraphic Analysis and the resulting diagram

The stratigraphic analysis is a non-destructive method based on archaeology that illustrates the relationships and sequences of the stratigraphic layers of excavations by listing all their constituting elements, to be later represented in a stratigraphic diagram. Although the constant improvements and applications of the stratigraphic analysis in several scientific fields since 1973, this study proposes specific modifications to the current method as well as an adaptation of the diagram scheme to each case study of conservation. The main goal of this study was the elaboration of a detailed and comprehensive diagram that encompasses the entire monument, rather than one for each individual section of the monument. The first step was the identification of the main stratigraphic units and their classification based on their primary function: structural or decorative. The second step concerned a simplification of the current relationships of the architectural units into three groups, according to their roles within the entire system to obtain a simpler stratigraphic sequence. The final step was the new incorporation of pathology-related information and the addition of the missing elements as a reconstruction process. These adjustments allowed the diagram to arrange all data gathered from heritage analysis and will permit historians, architects, archaeologists, and others to engage in a global reading of the built. The stratigraphic diagram will serve as a tool to visually represent the analysis and synthesis in a coded manner, which will be comprehensible to both the researchers and the scientific community.

Excitatory and inhibitory neuronal synapse unit: A novel recurrent cell for time series prediction

Time series prediction is one of the most challenging topics in real-world applications. Most studies undertook advanced mathematics and computer techniques above benchmark methods but often unable to guarantee forecasting models’ generalization and reliability. This paper proposed a biologically motivated recurrent unit based on neuronal synaptic activity mechanism and chaotic behaviors in deep learning domain called Excitatory and Inhibitory Neural Synapse unit (EINS). The major contribution of this research is to create a flexible and robust recurrent unit based on neuroscience theory. It conducted experiments on three real-world time series forecast examples including finance, household power, weather for generalization and robustness evaluation using eight state-of-the-art recurrent architecture units as baseline models to compare convergence rate and forecasting accuracy. Experimental results showed that EINS achieved satisfactory performances in time series prediction reliability and effectiveness.

A Novel Improved Variational Mode Decomposition-Temporal Convolutional Network-Gated Recurrent Unit with Multi-Head Attention Mechanism for Enhanced Photovoltaic Power Forecasting

Photovoltaic (PV) power forecasting plays a crucial role in optimizing renewable energy integration into the grid, necessitating accurate predictions to mitigate the inherent variability of solar energy generation. We propose a novel forecasting model that combines improved variational mode decomposition (IVMD) with the temporal convolutional network-gated recurrent unit (TCN-GRU) architecture, enriched with a multi-head attention mechanism. By focusing on four key environmental factors influencing PV output, the proposed IVMD-TCN-GRU framework targets a significant research gap in renewable energy forecasting methodologies. Initially, leveraging the sparrow search algorithm (SSA), we optimize the parameters of VMD, including the mode component K-value and penalty factor, based on the minimum envelope entropy principle. The optimized VMD then decomposes PV power, while the TCN-GRU model harnesses TCN’s proficiency in learning local temporal features and GRU’s capability in rapidly modeling sequence data, while leveraging multi-head attention to better utilize the global correlation information within sequence data. Through this design, the model adeptly captures the correlations within time series data, demonstrating superior performance in prediction tasks. Subsequently, the SSA is employed to optimize GRU parameters, and the decomposed PV power mode components and environmental feature attributes are inputted into the TCN-GRU neural network. This facilitates dynamic temporal modeling of multivariate feature sequences. Finally, the predicted values of each component are summed to realize PV power forecasting. Validation using real data from a PV station corroborates that the novel model demonstrates a substantial reduction in RMSE and MAE of up to 55.1% and 54.5%, respectively, particularly evident in instances of pronounced photovoltaic power fluctuations during inclement weather conditions. The proposed method exhibits marked improvements in accuracy compared to traditional PV power prediction methods, underscoring its significance in enhancing forecasting precision and ensuring the secure scheduling and stable operation of power systems.