Innovative Architecture Research Articles

Innovative lightweight deep learning architecture for enhanced rice pest identification

Pest detection is a crucial aspect of rice production. Accurate and timely identification of rice pests can assist farmers in taking prompt measures for control. To enhance the precision and real-time performance of rice pest detection, this paper introduces a novel YOLOv8-SCS architecture that integrates Space-to-Depth Convolution (SPD-Conv), Context Guided block (CG block), and Slide Loss. Initially, the original algorithm’s convolutional module is improved by introducing the SPD-Conv module, which reorganises the input channel dimensions into spatial dimensions, enabling the model to capture fine-grained pest features more efficiently while maintaining a lightweight model architecture. Subsequently, the CG block module is integrated into the CSPDarknet53 to 2-Stage FPN (C2f) structure, maintaining the models lightweight nature while enhancing its feature extraction capabilities. Finally, the Binary Cross-Entropy (BCE) is refined by incorporating the Slide Loss function, which encourages the model to focus more on challenging samples during training, thereby improving the model’s generalization across various samples. To validate the effectiveness of the improved algorithm, a series of experiments were conducted on a rice pest dataset. The results demonstrate that the proposed model outperforms the original YOLOv8 in rice pest detection, achieving an mAP of 87.9%, which is a 5.7% improvement over the original YOLOv8. The model also features a 44.1% reduction in parameter count and a decrease of 11.7 GFLOPs in computational requirements, meeting the demands for real-time detection.

Synthesis of Polyglycerol/PolyCaprolactone Nanocopolymers as Innovative Architectures for Drug Delivery

Synthesis of Polyglycerol/PolyCaprolactone Nanocopolymers as Innovative Architectures for Drug Delivery

Comprehensive review and study for the CMOS transformer tank voltage‐controlled oscillator

AbstractThis article comprehensively reviews and studies CMOS transformer tank voltage‐controlled oscillators (VCOs) designed to improve the frequency tuning range (FTR) and minimize phase noise (PN) across a broad frequency range, particularly for mm‐wave and fifth‐generation applications. The study focuses on enhancing the performance of the oscillators to meet the demanding requirements of these advanced applications. This study thoroughly investigates and compares various strategies to increase the FTR, minimize PN, and reduce power dissipation in transformer tank VCOs. The aim is to establish fundamental guidelines for implementing transformer‐based tank VCOs. The study comprehensively reviews and analyzes the overall architecture of transformer‐based tank VCOs, presenting a detailed examination of the methods employed to enhance VCO performance. Additionally, the study illustrates the impact of the coupling factor on the VCO's performance parameters, detailing and demonstrating the overall tank circuit quality factor (Q) and resonator coupling factor (κ). A thorough understanding of PN methodology and the limitations of FTR has led to the development of innovative architectures. These include transformer‐based capacitive feedback (FB) oscillators with resistively adjusted variable inductors, as well as class‐F and inverse class‐F VCOs. These novel architectures contribute significantly to the improvement of oscillator performance.

Downstream lingering attention transformer network (DsLATNet) for land use land cover classification: A bicolor deep learning framework

Land Use and Land Cover (LULC) classification is the process of locating and classifying regions of the Earth's surface (land cover) based on their physical attributes and human utilization (land use purpose). It is essential for mapping and monitoring changes in ecosystems, facilitating effective resource management and informed decision-making in land-use planning. Convolutional Neural Network (CNN’s) is incapable of capturing long-range dependencies. The effectiveness of transformers relies on extensive training datasets, yet many satellite datasets have comparatively limited sample sizes. Moreover, due to the numerous factors influencing visual prominence, it is challenging to obtain abundant features from a single-color space. To overcome these limitations, we introduce an innovative bicolor architecture Downstream Lingering Attention Transformer Network (DsLATNet). DsLATNet characterizes remote sensing images on two color spaces viz RGB and HSV by processing the information using Residual Network −50(ResNet-50) backbone network.Extracted features of backbone network are further refined to acquire detailed aware features through Downstream Lingering Feature Pyramid Network (DsLFPN).This is further integrated with a DuoTransformer that utilizes attention mechanisms to obtain long-range dependencies. Eventually, the features are aggregated using residual connections that are dilated in-depth multiple times to produce class labels. The efficacy of the algorithm is evaluated through standard classification accuracy metrics. Experimental results on both the Wuhan Dense Labeling Dataset (WHDLD) and the Gaofen Image Dataset (GID) exhibit that the suggested technique exceeds cutting-edge classification methods. WHDLD achieved an Overall Accuracy (OA) of 86.34 %, Average Accuracy (AA) of 75.62 %, and Kappa coefficient (K) of 80.75, while the GID exhibited superior performance with an OA of 86.43 %, AA of 88.13 %, and K of 81.25.

(Invited) Unlocking Long-Duration Energy Storage with Unitized Reversible Fuel Cells

As the proportion of energy supplied by renewable sources continually rises, the intermittent nature of these sources necessitates efficient energy storage to meet the demand of the grid. Hydrogen is a promising energy carrier for grid-scale storage due to its capacity for months-long storage without energy loss or self-discharge.(1) Reversible fuel cells generate power for the grid in fuel cell (FC) mode during periods when renewable power production is lower than demand, utilizing stored H2 generated by water electrolysis (WE) during periods when renewable power production exceeds demand from the grid. Unitized reversible fuel cells (URFCs) efficiently integrate WE and FC operation in a single electrochemical device, potentially reducing costs and providing higher energy density. In constant gas URFC operational configuration, both the oxygen evolution reaction (OER) for WE and the oxygen reduction reaction (ORR) for FC occur within the same electrode.(2) The contrasting electrode requirements for FC and WE have constrained the round-trip efficiency (RTE) of URFCs, highlighting the crucial need to improve RTE for greater competitiveness and widespread commercial adoption. This presentation will discuss the results from development of novel bifunctional catalysts, electrode architectures, and porous transport layers (PTLs) aimed at enhancing efficiency and durability of URFCs. This talk will present the development of bifunctional catalysts and their supports to boost OER and ORR performance. Additionally, we will present insights into innovative electrode architecture designed for optimal utilization of the catalysts, thereby maximizing efficiency. The development of specialized PTLs featuring amphiphilic wettability and intricately engineered microporous layers will be presented emphasizing their role in increasing the performance and durability of URFCs. Acknowledgement This research is supported by the Laboratory Directed Research and Development (LDRD) program at Los Alamos National Laboratory (LANL) under project 20240061DR, and by the U.S. Department of Energy Hydrogen and Fuel Cell Technologies Office (DOE-HFTO) with support from technology manager Donna Ho. References Hydrogen Council, Hydrogen scaling up: a sustainable pathway for the global energy transition, in (2017).S. Komini Babu et al., Advanced Energy Materials, 13, 2203952 (2023).

Coupling Dynamics and Linear Polarization Phenomena in Codirectional Polariton Waveguide Couplers

AbstractIn this work, the previous studies investigating the potential of integrated devices using polariton waveguides that form codirectional couplers is extended. For suitable coupler parameters, a transfer of condensates between the arms of the coupler occurs leading to the observation of previously predicted Josephson‐like oscillations. The ability to tune the periodicity of these oscillations opens the way to the design of polaritonic circuits in which the directionality of the signal towards the output terminals can be controlled, for a fixed separation between the arms, varying the coupling length. The response of the devices to linearly polarized excitation is also investigated, delving into the dynamics of linear polarization at the output terminals of long couplers, providing valuable insights into their potential applications, including polariton switches and logic gates with efficient operation. The results are supported by numerical simulations based on the generalized Gross–Pitaevskii equation describing the dynamics of coherent polaritons in spatially non‐uniform systems, reproducing the polariton distribution on the output terminals and an oscillatory dependence of the corresponding emission as a function of the polarization azimuth. How the coupling and the controllable polarization degree of freedom in polariton couplers opens avenues for innovative optical architectures and functionalities is shown here.

MARes-Net: multi-scale attention residual network for jaw cyst image segmentation.

Jaw cyst is a fluid-containing cystic lesion that can occur in any part of the jaw and cause facial swelling, dental lesions, jaw fractures, and other associated issues. Due to the diversity and complexity of jaw images, existing deep-learning methods still have challenges in segmentation. To this end, we propose MARes-Net, an innovative multi-scale attentional residual network architecture. Firstly, the residual connection is used to optimize the encoder-decoder process, which effectively solves the gradient disappearance problem and improves the training efficiency and optimization ability. Secondly, the scale-aware feature extraction module (SFEM) significantly enhances the network's perceptual abilities by extending its receptive field across various scales, spaces, and channel dimensions. Thirdly, the multi-scale compression excitation module (MCEM) compresses and excites the feature map, and combines it with contextual information to obtain better model performance capabilities. Furthermore, the introduction of the attention gate module marks a significant advancement in refining the feature map output. Finally, rigorous experimentation conducted on the original jaw cyst dataset provided by Quzhou People's Hospital to verify the validity of MARes-Net architecture. The experimental data showed that precision, recall, IoU and F1-score of MARes-Net reached 93.84%, 93.70%, 86.17%, and 93.21%, respectively. Compared with existing models, our MARes-Net shows its unparalleled capabilities in accurately delineating and localizing anatomical structures in the jaw cyst image segmentation.

Enhancing SFC Placement with Parallelized Functions in MEC Using Deep Reinforcement Learning

This study presents an innovative architecture leveraging Deep Reinforcement Learning (DRL) to optimize online service delivery in Mobile Edge Computing (MEC) networks. Our proposed method, the DRL-based dynamic Service Function Chain (SFC) placement with parallelized Virtual Network Functions (VNFs), aims to maximize Long-Term Cumulative Reward (LTCR). By parallelizing VNFs, our approach achieves computational acceleration, enhancing the efficiency of online service provisioning. Furthermore, we enhance future request processing capabilities by extracting the distribution of initialized VNFs. Through comprehensive simulations, our proposed architecture demonstrates a notable improvement, with approximately a 9.5% increase in the average number of accepted requests.

Study of readout system for 3He linear position sensitive detectors

This study introduces a readout system for 3He linear position-sensitive detectors (LPSDs), tailored for the Small Angle Neutron Scattering (SANS) spectrometer of the Compact Pulsed Hadron Source (CPHS) at Tsinghua University. Central to this system is the employment of a trapezoidal filter as the primary signal shaper, selected for its enhanced noise filtering capabilities and efficacy in mitigating ballistic deficit. Additionally, the integration of a digital CR filter aims to counteract baseline drift within trigger signals. This innovative architecture underwent validation via computer programs using experimental data, demonstrating superior performance in spatial resolution compared to conventional Gaussian filter-based approaches. Meanwhile, the improvement in spatial resolution comes with a narrower pulse width, promising a higher throughput at high count rates. When applied to a 16-tube detector module, the system achieved an average position resolution of 3.63 mm (FWHM) at the central region and 4.03 mm at the end regions. Furthermore, with a pusle width of 0.6 μs, the 96-tube detector array is promised to sustain a count rate over 1 Mcps.

Towards Millions of Database Transmission Services in the Cloud

Alibaba relies on its robust database infrastructure to facilitate realtime data access and ensure business continuity despite regional disruptions. To address these operational imperatives, Alibaba developed the Data Transmission Service (DTS), which has become critical for internal applications and public cloud services alike. This paper presents a comprehensive study of the architectural innovations, resource scheduling mechanisms, and performance optimization strategies that have been implemented within DTS to tackle the significant challenges of cross-network, heterogeneous data transmission in a cost-effective manner. We explore the novel Any-to-Any (A2A) architecture, which simplifies the complexity of data paths between diverse databases and mitigates network connectivity issues, thereby significantly reducing development overhead. Additionally, we examine a dynamic network bandwidth scheduling algorithm that effectively maintains Service-Level Objectives (SLOs), complemented by a serverless mechanism that ensures efficient resource utilization. Furthermore, DTS utilizes advanced strategies such as transaction dependency tracking, hot data consolidation, and batching to enhance synchronization performance and efficiency. DTS has distilled the lessons learned from years of serving our customer base and currently supports nearly 1 million public cloud instances annually. Our evaluation results show that DTS can effectively and efficiently handle real-time data transmission in both experimental and production environments.

From Novelty to Architectural Innovation: The Evolution and Scalability of Business Models

From Novelty to Architectural Innovation: The Evolution and Scalability of Business Models

Research on the performance of hybrid vision models based on ViT

ViT is a model proposed by the Google team in 2020 to apply a transformer in image classification, although it is not the first paper to apply a transformer in visual tasks, because of its model is simple and effective, and scalable, it has become a milestone work in the application of transformer in CV field, and has also triggered the subsequent related research. The core conclusion of the original ViT paper is that when there is enough data for pre-training, ViT outperforms CNN, breaks through the limitation of the transformer's lack of inductive bias, and can achieve better migration results in downstream tasks. However, when the training dataset is not large enough, ViT usually performs worse than ResNets of the same size, because Transformer lacks inductive bias, a kind of a priori knowledge, assumptions made in advance, compared with CNN. Through its innovative architecture and powerful performance, the visual representation transform (ViT) model continues to advance the field of computer vision, while facing some challenges and room for improvement. With the deepening of research and the continuous development of technology, ViT is expected to play a greater role in more practical applications. The article aims to explore the advantages and applicability of the ViT model and tries to construct a hybrid visual model to improve its generalization ability for different types of datasets, demonstrating the hybrid model's significance in improving the performance of the ViT model.

Efficient state of charge estimation of lithium-ion batteries in electric vehicles using evolutionary intelligence-assisted GLA–CNN–Bi-LSTM deep learning model

The battery's performance heavily influences the safety, dependability, and operational efficiency of electric vehicles (EVs). This paper introduces an innovative hybrid deep learning architecture that dramatically enhances the estimation of the state of charge (SoC) of lithium-ion (Li-ion) batteries, crucial for efficient EV operation. Our model uniquely integrates a convolutional neural network (CNN) with bidirectional long short-term memory (Bi-LSTM), optimized through evolutionary intelligence, enabling an advanced level of precision in SoC estimation. A novel aspect of this work is the application of the Group Learning Algorithm (GLA) to tune the hyperparameters of the CNN–Bi-LSTM network meticulously. This approach not only refines the model's accuracy but also significantly enhances its efficiency by optimizing each parameter to best capture and integrate both spatial and temporal information from the battery data. This is in stark contrast to conventional models that typically focus on either spatial or temporal data, but not both effectively. The model's robustness is further demonstrated through its training across six diverse datasets that represent a range of EV discharge profiles, including the Highway Fuel Economy Test (HWFET), the US06 test, the Beijing Dynamic Stress Test (BJDST), the dynamic stress test (DST), the federal urban driving schedule (FUDS), and the urban development driving schedule (UDDS). These tests are crucial for ensuring that the model can perform under various real-world conditions. Experimentally, our hybrid model not only surpasses the performance of existing LSTM and CNN frameworks in tracking SoC estimation but also achieves an impressively quick convergence to true SoC values, maintaining an average root mean square error (RMSE) of less than 1 %. Furthermore, the experimental outcomes suggest that this new deep learning methodology outstrips conventional approaches in both convergence speed and estimation accuracy, thus promising to significantly enhance battery life and overall EV efficiency.

Innovative Sustainable Architecture: A Lesson Learned from Amphibious House in the UK

Innovative Sustainable Architecture: A Lesson Learned from Amphibious House in the UK

CFD modeling of braid-reinforced hollow fiber membranes for efficient water and bovine serum albumin (BSA) separation

This study presents a comprehensive investigation into the performance optimization of a braid-reinforced hollow fiber membrane (BHFM) system for protein separation from milk wastewater. Extensive CFD simulations were conducted to examine the effects of various effective parameters on system performance. The results reveal that at a feed protein concentration of 100 mg/l, the average water flux increases from 149.16 to 453.68 l/m2.h as the inlet feed pressure rises from 100 to 300 kPa. Similarly, for a feed protein concentration of 300 mg/l, the average flux rises from 98.69 to 299.84 l/m2.h. However, higher feed concentrations lead to increased fouling and lower water fluxes. Additionally, an increase in feed pressure results in a decrease in the protein rejection rate, particularly in the initial stages of separation. The study highlights that variations in membrane porosity and thickness significantly affect both protein rejection rate and water flux. Increasing membrane porosity from 0.2 to 0.8 leads to a decrease in protein rejection rate from 94.57 % to 78.47 % at an inlet pressure of 100 kPa, and from 93.33 % to 73.43 % at 300 kPa. Similarly, increasing membrane thickness from 0.1 to 1.0 mm decreases the time-averaged water flux from 105.76 to 8.64 l/m2.h at 100 kPa, and from 321.39 to 25.99 l/m2.h at 300 kPa. However, increasing membrane thickness also increases the protein rejection rate. The study's novelty lies in the quantitative results obtained through CFD simulations, providing valuable insights into BHFM system performance. Future research should focus on exploring innovative braid architectures and material selection to enhance the mechanical stability and durability of braid-reinforced structures under various operational conditions.

Energy efficiency research of propulsion system for series-parallel hybridization of amphibious vehicles

The series-parallel hybrid system has attracted much attention from scholars for its effective integration of the power advantages and operating characteristics of different power sources, which is influenced by international emission regulations, energy-saving and emission reduction policies. As such, a series-parallel hybrid powertrain is introduced to the amphibious vehicle, and an innovative powertrain topology architecture is proposed. Meanwhile, the operation mode and energy efficiency characteristics are investigated during the working process. Firstly, the energy flow simulation model of a series-parallel gas-electric hybrid propulsion system is constructed using a modular modeling approach. Secondly, four operating modes, namely mechanical propulsion, electric propulsion, hybrid propulsion and charging mode, were formulated due to the fact that the propulsion system has multiple forms of power sources in the form of natural gas engine and reversible motor. Meanwhile, the energy flow states were investigated under different operating modes. Meanwhile, a comprehensive investigation of the energy efficiency associated with propulsion, storage and start-up energy was conducted for each specific mode. The results of the research indicated that the energy efficiency of the electric propulsion mode can reach up to 35.15 %, which is the gain from the wide operating range of the motor's high efficiency. The hybrid propulsion mode can obtain the highest energy efficiency of 35.88 %, which fully demonstrates the advantages of coordinating and complementing the two power sources, the natural gas engine and the reversible electric motor. This investigation also provides theoretical and empirical support for optimizing energy matching and formulating energy management strategies.

Education, innovation and research in wooden architecture and construction in the Alps

The essay explores the importance of teaching and research in the building cultures of specific territories, with a particular focus on the use of wood in construction. Through interviews with experts from research centres and universities, such as Andreja Kutnar from InnoRenew CoE in Slovenia, Frédéric Pichelin from the Bern University of Applied Sciences in Switzerland, and Florian Court from the Haute École du Bois et de la Forêt association in France, the essay highlights how such institutions are promoting innovation and sustainable development in the wooden construction sector. Kutnar emphasises the importance of interdisciplinary and holistic research to develop renewable materials and improve the quality of the built environment. Pichelin discusses how innovation can emerge from the integration of new technologies and biobased materials, while Court addresses the challenge of industrialising wood production processes while preserving unique local characteristics. The essay also examines future challenges for wooden architecture in the Alps, such as climate change and the need for sustainable resources, highlighting projects like VETA/NOVA developed by the Bern University of Applied Sciences, which aims to develop guidelines for the restoration of historic wooden buildings. Finally, it discusses the possibility of translating global prerogatives into a local context, influencing the contemporary architectural identity of the Alps and its multidimensional impact on the territory.

Architecture and local resources: project experiences in Vorarlberg

The Land of Vorarlberg is Austria’s second smallest but most densely populated province after Vienna. Known for its innovative timber architecture, Vorarlberg’s approach integrates ecological, economic, functional, and aesthetic considerations, as championed by Wolfgang Ritsch of the Vorarlberger Architektur Institut (VAI). The region’s timber industry has faced challenges, including local provincialism and modernist material preferences, but gained prominence in the 1990s with the establishment of the Baukünstler group and the Vorarlberger Bauschule, recognised for their contributions to timber construction. Key to this development is Vorarlberg’s significant woodland, covering one-third of its area, and its commitment to sustainable forest management. Regional timber architecture aligns with the principles of appropriate technology, emphasizing eco-responsible, community-driven innovations. Vorarlberg’s timber architecture exemplifies a collaborative effort between designers and craftsmen, focusing on the use of local resources and high construction quality. The essay represents the introduction of notable projects including multifunctional public buildings, schools, offices, and residential complexes, reflecting the region’s commitment to sustainable development and architectural excellence. This ongoing innovation in wood-integrated solutions reinforces Vorarlberg as a hub for eco-friendly and culturally resonant architectural practices.

DeMambaNet: Deformable Convolution and Mamba Integration Network for High-Precision Segmentation of Ambiguously Defined Dental Radicular Boundaries.

The incorporation of automatic segmentation methodologies into dental X-ray images refined the paradigms of clinical diagnostics and therapeutic planning by facilitating meticulous, pixel-level articulation of both dental structures and proximate tissues. This underpins the pillars of early pathological detection and meticulous disease progression monitoring. Nonetheless, conventional segmentation frameworks often encounter significant setbacks attributable to the intrinsic limitations of X-ray imaging, including compromised image fidelity, obscured delineation of structural boundaries, and the intricate anatomical structures of dental constituents such as pulp, enamel, and dentin. To surmount these impediments, we propose the Deformable Convolution and Mamba Integration Network, an innovative 2D dental X-ray image segmentation architecture, which amalgamates a Coalescent Structural Deformable Encoder, a Cognitively-Optimized Semantic Enhance Module, and a Hierarchical Convergence Decoder. Collectively, these components bolster the management of multi-scale global features, fortify the stability of feature representation, and refine the amalgamation of feature vectors. A comparative assessment against 14 baselines underscores its efficacy, registering a 0.95% enhancement in the Dice Coefficient and a diminution of the 95th percentile Hausdorff Distance to 7.494.

Development of conductive carbon-based interdigitated electrode array integrated with microfluidic channel for blood glucose sensing

PurposeThis study aims to provide a microfluidic blood glucose sensing platform based on integrated interdigitated electrode arrays (IDEAs) on a flexible quartz glass substrate, adhering closely to pertinent electrochemical characterizations.Design/methodology/approachSensors are the key elements of the modern electronics era through which all the possible physical quantities can be detected and converted into their equivalent electrical form and processed further. But to make the sensing environment better, various types of innovative architectures are being developed nowadays and among them interdigitated electrodes are quite remarkable in terms of their sensing capability. They are a well-qualified candidate in the field of gas sensing and biosensing, but even their sensitivities are getting saturated due to their physical dimensions. Most of the thin film IDEAs fabricated by conventional optical lithographic techniques do not possess a high surface-to-volume ratio to detect the target specified and that reduces their sensitivity factor. In this context, a classic conductive carbon-based highly sensitive three dimensional (3D) IDEA-enabled biosensing system has been conceived on a transparent and flexible substrate to measure the amount of glucose concentration present in human blood. 3D IDEA possesses a way better capacitive sensing behavior compared to conventional thin film microcapacitive electrodes. To transmit the target biological analyte sample property for the detection purpose to the interdigitated array-based sensing platform, the design of a microfluidic channel is initiated on the same substrate. The complex 3D Inter Digital array structure improves the overall capacitance of the entire sensing platform and the reactive surface area as well. The manufactured integrated device displays a decent value of sensitivity in the order of 5.6 µA mM−1 cm−2.FindingsDevelopment of a low-cost array-based integrated and highly flexible microfluidic biochip to extract the quantity of glucose present in human blood.Originality/valuePotential future research opportunities in the realm of integrated miniaturized, low-cost smart biosensing systems may arise from this study.