As an important part of the construction industry, rural residential buildings are characterized by low energy utilization, unreasonable structures and low consumption levels, and it is particularly important to study their low-carbon transformation and evaluation system. In view of the many low-carbon transformation needs of rural residential buildings, the existing research results were analyzed in depth, and the coefficient of variation method was used to identify the important factors affecting the low-carbon transformation of rural residential buildings, and the evaluation system of rural residential buildings’ low-carbon transformation was determined by Analytic Hierarchy Process (APH), and the system was used in a rural residential building low-carbon evaluation study. The results show that the influence of “energy use”, “envelope structure” and “economic factors” on the decarbonization of buildings is obvious, with the weights of 36.4%, 24.5% and 19.5% respectively. Among the secondary indicators, “clean energy utilization”, “electricity consumption”, “external wall insulation system” and “window performance” are the most important factors in reducing carbon emissions in rural areas. The most critical influencing factors for the low carbonization level of clean energy in rural residential buildings are “window performance”. Finally, based on the constructed low carbonization evaluation system, we propose a targeted solution strategy to provide a theoretical basis for the establishment of an effective low carbonization evaluation system for clean energy in rural residential buildings.

The article discusses the interconnected fields of computing and machine learning, and their impact on various areas such as energy, economics, indoor positioning, and business. Computing provides the foundation for data processing and storage, while machine learning enables algorithms and models to learn from data and make predictions. These advancements have revolutionized how we approach complex problems and opened up new avenues for research and innovation. The article highlights the potential of computing and data science to solve complex problems and the importance of staying up-to-date with the latest developments.

A document by Gengrui Chen. Click on the document to view its contents.

Model ensemble is widely used in deep learning since it can balance the variance and bias of complex models. The mainstream model ensemble methods can be divided into “implicit” and “explicit”. The “implicit” method obtains different models by randomly inactivating the internal parameters in the complex structure of the deep learning model, and these models are integrated by sharing parameters. However, these methods lack flexibility because they can only ensemble homogeneous models with the similar structure. While the “explicit” ensemble method can fuse completely different heterogeneous model structures, which significantly enhances the flexibility of model selection and makes it possible to integrate more models with entirely different perspectives. However, the explicit ensemble will face the challenge of averaging the outputs, leading to a chaotic result. To this end, researchers further proposed using knowledge distillation and adversarial learning technologies to perform a nonlinear combination of multiple heterogeneous models to obtain better ensemble performance, however these require significant modifications to the training or testing procedure and are computationally expensive compared to simply averaging. In this paper, based on the linear combination assumption, we propose an interpretable ensemble method for averaging model results which is simple to implement, and conducting experiments on the representation learning tasks of Computer Vision(CV) and Natural Language Processing(NLP). The results show that our method is superior to direct averaging results while retaining the practicality of direct averaging.

Anemia is one of the global public health challenges that particularly affect children and pregnant women. A study by WHO indicates that 42% of children below 6 years and 40% of pregnant women worldwide are anemic. This affects the world’s total population by 33%, due to the cause of iron deficiency. The non-invasive technique, such as the use of machine learning algorithms, is one of the methods used in the diagnosing or detection of clinical diseases, which anemia detection cannot be overlooked in recent days. In this study, machine learning algorithms were used to detect iron-deficiency anemia with the application of Naïve Bayes, CNN, SVM, k-NN, and Decision Tree. This enabled us to compare the conjunctiva of the eyes, the palpable palm, and the colour of the fingernail images to justify which of them has a higher accuracy for detecting anemia in children. The technique utilized in this study was categorized into three different stages: collecting of datasets (conjunctiva of the eyes, fingernails and the palpable palm images), preprocessing the images; image extraction, segmentation of the Region of Interest of the images, obtained each component of the CIE L*a*b* colour space (CIELAB). The models were then developed for the detection of anemia using various algorithms. The CNN had an accuracy of 99.12% in the detection of anemia, followed by the Naïve Bayes with an accuracy of 98.96%, while Decision Tree and k-NN had 98.29% and 98.92% accuracy respectively. However, the SVM had the least accuracy of 95.4% on the palpable palm. The performance of the models justifies that the non-invasive approach is an effective mechanism for anemia detection. Keywords: Iron deficiency, anemia, non-invasive, machine learning, data augmentation, algorithms, region of interest.

Error Augmentation training using a robotic interface is thought to promote motor recovery by enhancing proprioceptive feedback, which motivates and challenges patients to optimize their performance during training. Here, we investigated the effectiveness of robotic Error Augmentation training on motor recovery after a stroke, compared to standard robotic training in a null field. Post-stroke patients were randomly assigned to one of two groups: a study group (n=9) that was trained on a 3D robotic system applying Error Augmentation forces, and a control group (n=7) that carried out the same protocol in null field conditions. The robotic rehabilitation intervention was applied in addition to the standard rehabilitation protocol of the rehabilitation center. Error Augmentation training increased clinical scores compared to standard robotic training by 266% on the Motor Assessment Scale, and 88% on the Fugl-Meyer scale. The Motor Assessment Scale scores were significantly correlated with the Fugl-Meyer scores (p=0.03, r=0.541). There were more movement errors on the initial trials of the game sequence using the DeXtreme robotic device with Error Augmentation compared to trials with no force field. This difference vanished however after 10 trials. Error Augmentation training decreased the number of movement units and jerkiness compared to the control treatment. These findings suggest that Error Augmentation training may enhance motor performance possibly through motor adaptation.

A complete shape factor investigation of water-based mixture type hybrid nano-fluid in a permeable boundary with the impact of magnetic field, thick dissemination, and warm radiation is presented in this article. A computational convection analysis of an inverted semi vertical cone with a porous surface in the form of S i O 2 / w a t e r nano-fluid and M O S 2 − S i O 2 / w a t e r hybrid nano-fluid transport is developed. The system of differential equations is presented and resolved numerically by the Lobatto IIIA method. The temperature distributions and fluid velocity are studied along with the coefficient of skin friction and the nusselt number, taking into account the form of distinct nano-particles. The flow problem’s results are approximated by using several embedding variables. Tables and graphs are constructed for a variety of scenarios including maximum residual error, mesh points, and nusselt numbers. We conclude that boundary film thickness reduces and the fluid flow is resisted by magnetic field presence. Fluid flow slows down as λ increases, and this reduction is more evident in nanofluids than in hybrid nanofluids. With an increment in S, velocity drops. A detailed analysis of the proposed ordinary differential equations, boundary conditions, and numerical data of skin friction is given both in tabular and graphical forms. Additionally, it is observed that the fluid flow slows down more for the hybrid nanofluid than for the SiO2 /water nanofluid. Additionally, it is clear that the temperature increase for the SiO2 /water nanofluid is substantially greater. The authors deduce that the existence of a magnetic field resists fluid flow for hybrid nanofluid forms and decreases the thickness of the viscous boundary layer.

Internet of Things (IOT) related and IOT enabled education and research are becoming more prevalent in today’s academic environment. One of the challenges faced by educators and researchers is the availability of resources to support lab based, hands on learning and research projects. Existing resources are often highly customized, balkanized, and difficult to manage. For new technology areas such as IOT, these environments rarely provide opportunities for resource sharing and interdisciplinary collaboration. In this paper we describe the development of a multi-tenant, IOT centric platform, the Environmental Sensing Data Network (ESDN) meant to support plug and play sensor deployment and the tools necessary for data management. The platform turns the University campus into an open platform that can be used to easily set up IOT based labs for education, research, campus infrastructure management and community applications innovation. The paper covers the development of a sensor augmentation module facilitating easy deployment and management. The operational tools and services integration are also described. We believe the platform is an important step in facilitating distance and in person education and research. ESDN is rapidly expanding beyond the multi-site University campus supporting a growing number of research and community projects.

Cancer diagnosis using gene expression data is significant research for facilitating early treatment and prevention of cancer. The classification of gene expression data is challenging due to its high dimensionality and smaller number of samples that renders classification a difficult task. Creation of well-defined class boundaries is the aim of every classification algorithm. The Fuzzy min-max (FMM) neural network classifier is known to create good decision boundaries using hyperboxes constructed for each class. In this paper, we explore the General Fuzzy min-max (GFMM) and Enhanced Fuzzy min-max (EFMM) neural network architectures for the classification of lung cancer subtypes from microarray gene expression data. Both GFMM and EFMM are advanced versions of Simpson’s FMM neural network classifier. The GFMM is extremely efficient because it involves very simple operations for hyperbox manipulation, and can handle both labeled and unlabeled data. On the other hand, EFMM proposes three heuristic rules related to hyperbox expansion, contraction and the overlap test, which enhances the learning algorithm. We perform the classification of gene expression data using these two algorithms, then we analyze the performance by visualizing the hyperboxes obtained after training, and compare the accuracies of these classifiers. LASSO is used for selecting the important genes from the high-dimensional gene expression data. After the analysis of the results, we observe that EFMM with LASSO gives the best performance as compared to GFMM, FMM and other machine learning algorithms.

Zk-SNARK unleashes the great potential of ZKP (zero-knowledge proof) in the blockchain, distributed storage, etc. However, the proof-generation of zk-SNARK is excessively time intensive, making it a challenge to deploy a high-performance zk-SNARK in most real applications. As a result, NTT (Number Theoretic Transform), one of the most time-consuming parts in proof-generation, needs to be accelerated significantly. To address this issue, we propose a novel and efficient “data reordering” technique to enable a highly pipelined architecture, on which an FPGA-based hardware accelerator is designed to support the large-bitwidth and large-scale NTT tasks in zk-SNARK. Our architecture achieves a two-level pipeline: 1) the top-level pipeline is achieved among smaller NTT sub-tasks, which are decomposed from a large-scale NTT task; 2) the bottom-level pipeline is achieved in each sub-task, among butterfly operations with different step sizes. This architecture can effectively reduce the data dependency and memory access requirements, meanwhile, can be flexibly scaled to different scales of FPGAs. To balance computing efficiency and flexibility, the OpenCL equipped with HLS is used to implement the heterogeneous acceleration system. We prototype the accelerator on the AMD-Xilinx Alveo U50 card (UltraScale+ XCU50 FPGA). The evaluation results show that 1) our accelerator shows high scalability for different scales of FPGAs with a stable performance improvement; 2) it performs 1.95× faster than the one in PipeZK; 3) and it achieves 27.98×, 1.74× speedup and 6.9×, 6× energy efficiency improvement than AMD Ryzen 9 5900X single core and 12 cores respectively when integrated into the well-known ZKP open-source project, Bellman.

A computational study of Non-Newtonian (Casson) free convective MHD unsteady fluid flow has been highlighted in this article with mass and heat transit property through a vertical infinite porous plate. A sinusoidal boundary conditions have been considered as well as chemical reaction and thermal radiation. Using a collection of non-dimensional variables, the flow related equations are also turned into non-dimensional form. The EFDM algorithm is employed in order to arrive at a numerical solution via Compaq Visual Fortran 6.6a. The reliability of the numerical solution has been confirmed using stability testing and convergence analysis. The whole system is convergent at the value of and . A visual depiction of the impact of the pertinent factors on dimensionless velocity, temperature, and concentration profiles is displayed along with thorough explanations and graphical representation as well as tabular representation. Key finding of this work is that when the magnetic component is regarded in sinusoidal form, it greatly affects the heat transfer factors of Casson fluid and the heat rises as the results of heat source parameter, radiation parameter and Eckert number. It is also found that the Sherwood number is increased as the impact of chemical reaction parameter and the Lewis number, also the skin friction is decreased as the influence of porosity term got accelerated. As a last step in verifying the earlier study, the present results are contrasted with the results that were previously published.

Metasurfaces tuning is performed using different ways for wide range of applications. This study presents the design of thermally-tuned all-dielectric reconfigurable metasurface. A microfluidic channel, filled with different concentrations of tellurium – selenium (Te-Se) alloy, is added on the top of the elliptical dielectric resonator (EDR) unit cell of the considered metasurface. The electrical properties of used semiconductor alloy are varied in the range of 400°C to 700°C (steps size of 100°C). The impact of thermal tuning on the reflection and transmission characteristics of the designed metasurface is analyzed in the frequency range 20-30 GHz using COMSOL Multiphysics. Obtained results demonstrated that the realized metasurface exhibits reconfigurable behavior in terms of variations in the reflection and transmission characteristics with a change in either temperature or concentrations of selenium and tellurium. The wider bands with high reflection and low transmission frequency bands are obtained with lower concentrations of selenium and tellurium for all operating temperatures.

With the quick development of flexible memory electronics, multifunctional organic materials have been the necessary for fabricating electronics. In this work, the highly transparent and flexible electrode was successfully prepared by coating the high-performance silver nanowires (AgNWs) onto the colorless polyimide (PI) substrate. The prepared flexible PI-AgNWs electrodes exhibited a low sheet resistance of 15 Ω/sq with the high transparency of 68 % at the wavelength of 400 nm. A novel kind of polyimide TPC6FPI was successfully synthesized and characterized with excellent thermal stabilities and high glass transition temperature (Tg) above 250 °C. Furthermore, a kind of flexible transparent PI-AgNWs/ TPC6FPI/Al resistive memory device was prepared and exhibited excellent SRAM switching behavior with the threshold voltage of around 2.1V and the ON/OFF current ratio of ~10-4, which indicated that multifunctional PI-based memory device showed the potential to the wearable devices.

To improve the thermal properties of thermite safely and stably, electrostatic spraying was used to prepare the Al/MoO3 thermite. The Al/MoO3 thermites were detected and characterized by scanning electron microscopy (SEM) and X-ray diffraction (XRD), and thermal decomposition experiments were carried out by differential scanning calorimetry (DSC). The heat release of the Al/MoO3 thermite prepared by the electrostatic spray (1044 J·g-1) is significantly higher than that of the thermite prepared by the ultrasonic (692 J·g-1), which is due to more uniform dispersion between Al and MoO3. The initial reaction temperature and activation energy (Ea) of the former keep it steady. Electrostatic spray ensures the safety and stability of the Al/MoO3 thermite. This study provides a new idea for safely and stably improving the thermal properties of thermite by enhancing surface homogenization, which is of great significance for practical applications.

In this paper, the leader-follower architecture is constructed by combining intermittent-influence leaders with a signed social network. Unlike a typical network with leaders where leaders are supposed to continuously influence followers, in this article, the leaders intermittently influence followers. Furthermore, the number of influences is limited. We focus on how intermittent-influence leaders impact the evolution of followers' opinions. The relationship between followers' opinions and the number of leader broadcasts is analyzed in detail. Then, the number of broadcasts is regarded as the cost, and the changing trend of the revenue per broadcast is obtained. The results show that as the number of broadcasts increases, the revenue per broadcast decreases gradually. Finally, the concept of assimilation is introduced to weigh the costs and benefits, and the minimum number of broadcasts required for the leader to assimilate the followers is derived. Two examples are given to demonstrate the validity of the main conclusions.

Full-color 3D printing technology has become more and more popular for industrial manufacturing applications. The voxelization of color 3D models and the expression of appearance colors become more and more important. To illustrate the complex appearance and internal characteristics of 3D models, we present a voxelization algorithm of the color 3D models for color 3D printing in this paper. Specifically, we use the 3D model of 3MF with color appearance features for surface voxelization first. Then, we propose an improved scan line filling algorithm to realize the voxelization of the interior of the model and obtain the color voxel model. Finally, to improve the high quality of color printing, we propose a surface color inward diffusion algorithm, which can make the surface color of the voxel model diffuse inward and further express the surface color of the voxel model. The experimental results show our proposed method can obtain an effective color voxel model and accurately. express the color of the voxel model surface, which can enhance the color effect of the color model and realize the control of voxels.

In Sub-Saharan Africa, Professionals visually analyse the plants by looking for disease markers on the leaves to diagnose cassava infections, however, this method is extremely subjective. Automating the identification and classification of crop diseases may improve the accuracy of professional disease diagnosis and enable farmers in remote areas to monitor their crops without the assistance of experts. Algorithms for machine learning have been used in the early detection and classification of crop diseases. Motivated by the current developments in the field of Gaussian Processes, this study proposes to integrate the transfer learning approach with a deep Gaussian convolutional neural network model (DGCNN) for the detection and classification of cassava diseases. During this study, we used MobileNet V2 and VGG16 pre-trained transfer learning models and a hybrid kernel. Experiments with MobileNet V2 and a hybrid kernel revealed an accuracy of 90.11%. Also, experiments with VGG16 and a hybrid kernel revealed an accuracy of 88.63%. The major limitation of this study was computing resources since we used an ordinary computer in all our experiments. In our future work, we will experiment with the three kernel functions used in this study with kernel algorithms such as support vector machines and compare the results with those obtained during this study.

Geopolymer concrete shares similar mechanical properties with ordinary Portland cement (OPC) concrete, and is even provided with better performances in high temperature and high corrosion circumstances. However, geopolymer binder is still subject to the disadvantages of large shrinkage and high brittleness, which greatly limit its application. Fiber reinforcement is widely used in various geopolymer systems to overcome the brittleness issue, but retains the high strength. Over the past 10 years, a significant advance has been made in the research of fiber reinforced geopolymers in terms of toughening efficiency and durability improvement. This paper, as a mini review, focuses on three types of fibers, i.e., inorganic fiber, natural fiber and synthetic fiber, in geopolymers, and their specific effects on compressive, flexural and tensile strengths, fractural toughness, shear strength and durability including shrinkage, chemical and freezing-thaw resistances. The recement understanding of bonding mechanism and fiber-geopolymer interface are also discussed, and knowledge gaps and future work challenges are correspondingly pointed out.