Data Size Increases Research Articles

Merit: multi-level graph embedding refinement framework for large-scale graph

The development of the Internet and big data has led to the emergence of graphs as an important data representation structure in various real-world scenarios. However, as data size increases, computational complexity and memory requirements pose significant challenges for graph embedding. To address this challenge, this paper proposes a multilevel embedding refinement framework (MERIT) based on large-scale graphs, using spectral distance-constrained graph coarsening algorithms and an improved graph convolutional neural network model that addresses the over-smoothing problem by incorporating initial values and identity mapping. Experimental results on large-scale datasets demonstrate the effectiveness of MERIT, with an average AUROC score 8% higher than other baseline methods. Moreover, in a node classification task on a large-scale graph with 126,825 nodes and 22,412,658 edges, the framework improves embedding quality while enhancing the runtime by 25 times. The experimental findings highlight the superior efficiency and accuracy of the proposed approach compared to other graph embedding methods.

Social context‐aware macroscopic routing scheme for opportunistic network

AbstractOpportunistic networks (OppNets) have attracted widespread attention as wireless technologies have advanced. OppNets are widely used in delay‐tolerant applications because they route messages using a store‐carry‐forward mechanism. Recently, socially aware routing has been increasingly modeled for message dissemination in OppNets, where the message is routed selectively through cooperative nodes based on user interests; however, routing becomes extremely difficult as node density and data size increase. However, the current method fails to reduce data redundancy, message overhead, delay, and improve performance efficiency. To address the issues, this article proposes social context‐aware microscopic routing (SCAMR) for OppNets. SCAMR uses cluster‐based communication, novel social‐context association mapping, and an improved lost packet retrieval mechanism with minimal messaging overhead. In this work, the experiment was performed by considering three scenarios: varying node size, varying buffer size, and varying time‐to‐live size. The experimental results show that the SCAMR scheme improves delivery ratio by 71.25%, 67.87%, 69.18%, reduces delay by 33.93%, 26.68%, 35.36%, reduces the number of hop nodes (i.e., messaging overhead) by 77.84%, 71.53%, 76.04% over existing approaches namely, SCARF (SoCial‐Aware Reliable Forwarding Technique for Vehicular Communications), SRS (secure routing strategy), and EDT (effective data transmission) considering different scenarios, respectively.

P4SGD: Programmable Switch Enhanced Model-Parallel Training on Generalized Linear Models on Distributed FPGAs

Generalized linear models (GLMs) are a widely utilized family of machine learning models in real-world applications. As data size increases, it is essential to perform efficient distributed training for these models. However, existing systems for distributed training have a high cost for communication and often use large batch sizes to balance computation and communication, which negatively affects convergence. Therefore, we argue for an efficient distributed GLM training system that strives to achieve linear scalability, while keeping batch size reasonably low. As a start, we propose P4SGD, a distributed heterogeneous training system that efficiently trains GLMs through model parallelism between distributed FPGAs and through forward-communication-backward pipeline parallelism within an FPGA. Moreover, we propose a light-weight, latency-centric in-switch aggregation protocol to minimize the latency of the AllReduce operation between distributed FPGAs, powered by a programmable switch. As such, to our knowledge, P4SGD is the first solution that achieves almost linear scalability between distributed accelerators through model parallelism. We implement P4SGD on eight Xilinx U280 FPGAs and a Tofino P4 switch. Our experiments show P4SGD converges up to 6.5X faster than the state-of-the-art GPU counterpart.

A Deficiency of the Predict-Then-Optimize Framework: Decreased Decision Quality with Increased Data Size

This paper presents an analysis of the decision quality of the predict-then-optimize (PO) framework, an extensively used prescriptive analytics framework in uncertain optimization problems. Our primary aim is to investigate whether an increase in data size invariably leads to better decisions within the PO framework. We focus our analysis on two contextual stochastic optimization problems—one with a non-linear objective function and the other with a linear objective function—under the PO framework. The novelty of our work lies in uncovering a previously unknown relationship: the decision quality can deteriorate with increasing data size in the non-linear case and exhibit non-monotonic behavior in the linear case. These findings highlight a potential pitfall of the PO framework and constitute our main contribution to the field, offering invaluable insights for both researchers and practitioners.

RNdN: Fast Query Compilation for NVIDIA GPUs

GPU database systems are an effective solution to query optimization, particularly with compilation and data caching. They fall short, however, in end-to-end workloads, as existing compiler toolchains are too expensive for use with short-running queries. In this work, we define and evaluate a runtime-suitable query compilation pipeline for NVIDIA GPUs that extracts high performance with only minimal optimization. In particular, our balanced approach successfully trades minor slowdowns in execution for major speedups in compilation, even as data sizes increase. We demonstrate performance benefits compared to both CPU and GPU database systems using interpreters and compilers, extending query compilation for GPUs beyond cached use cases.

Application of internet of things data processing based on machine learning in community sports detection

Due to the rapid development of the IoT ecosystem, a large amount of IoT data has been generated. However, in the current development process of the Internet of Things, how to ensure data security is a key issue. Traditional IoT security protection is relatively basic. With the increase of data size, these facilities need to be further improved in terms of computing, information security, storage information, and stability. However, traditional security and privacy protection methods are often insufficient for large topology structures. Blockchain is based on distributed networks, characterized by monotonous and decentralized data. Blockchain also provides a new approach to encrypting IoT security information. This article introduces blockchain technology into the Internet of Things, which can prevent reliance on centralized servers and effectively protect internet data and information security. Then this article also analyzed the motion recognition and detection of community sports, as well as the widespread application of computer vision and machine learning technologies in various fields such as computer mining and data security. Motion detection technology is very important in the field of computer vision and has applications in various real-world scenarios. This article studies new ideas for information encryption in the computer Internet of Things, and also analyzes machine learning and motion detection systems and applies them to community sports, improving the development of community sports.

An efficient multiple-word embedding-based cross-domain feature extraction and aspect sentiment classification

As the size of the feature space and data size increases, it is difficult to find the essential key features for cross-domain classification problems. Traditional word embedding and feature selection models use limited-sized data and dimensions for feature ranking and classification processes. To overcome these challenges, a hybrid cross-domain classification (CDC) based feature selection has been proposed to improve the efficiency of aspect sentiment classification on large databases. Initially, the crawled data from the Amazon product dataset to the cloud web server. In the data filtering approach, each record is pre-processed for the noise removal process. In the CNN framework, the different word-embedded models are used to ensemble the features from the training data. Here, TF-ID+Word2Vec, Hybrid Glove, and Skipgram are used to select the relevant key features from the training data. On the local side, remote IoT-based devices data distributed processing. For in-depth analysis of the sentiments IOT sensors have been used for automation. The filtered key features are given to the proposed feature extraction measure to select the essential features for the cross-domain classification model, it classifies the positive or negative. The proposed CDC model achieves overall precision and recall is 0.96 and 0.97 respectively. Experimental results proved that the proposed cross-domain feature selection-based classification approach has a better overall true positive than the conventional approaches.

Real-time batch processing at a GPU-based edge with a passive optical network

In recent years, advances in deep learning technology have significantly improved the research and services relating to artificial intelligence. Real-time object recognition is an important technique in smart cities, in which low-cost network deployment and low-latency data transfer are key technologies. In this study, we focus on time- and wavelength-division multiplexed passive optical network (TWDM-PON)-based inference systems to deploy cost-efficient networks that accommodate several network cameras. A significant issue for a graphics processing unit (GPU)-based inference system via a TWDM-PON is the optimal allocation of the upstream wavelength and bandwidth to enable real-time inference. However, an increase in the batch size of the arrival data at the edge servers, thereby ensuring low-latency transmission, has not been considered previously. Therefore, this study proposes the concept of an inference system in which a large number of cameras periodically upload image data to a GPU-based server via the TWDM-PON. Moreover, we propose a cooperative wavelength and bandwidth allocation algorithm to ensure low-latency and time-synchronized data arrivals at the edge. The performance of the proposed scheme is verified through a computer simulation.

High-Dimensional Variable Selection for Quantile Regression Based on Variational Bayesian Method

The quantile regression model is widely used in variable relationship research of moderate sized data, due to its strong robustness and more comprehensive description of response variable characteristics. With the increase of data size and data dimensions, there have been some studies on high-dimensional quantile regression under the classical statistical framework, including a high-efficiency frequency perspective; however, this comes at the cost of randomness quantification, or the use of a lower efficiency Bayesian method based on MCMC sampling. To overcome these problems, we propose high-dimensional quantile regression with a spike-and-slab lasso penalty based on variational Bayesian (VBSSLQR), which can, not only improve the computational efficiency, but also measure the randomness via variational distributions. Simulation studies and real data analysis illustrated that the proposed VBSSLQR method was superior or equivalent to other quantile and nonquantile regression methods (including Bayesian and non-Bayesian methods), and its efficiency was higher than any other method.

Tumor-Net: convolutional neural network modeling for classifying brain tumors from MRI images

Abnormal brain tissue or cell growth is known as a brain tumor. One of the body's most intricate organs is the brain, where billions of cells work together. As a head tumor grows, the brain suffers damage due to its increasingly dense core. Magnetic resonance imaging, or MRI, is a type of medical imaging that enables radiologists to view the inside of body structures without the need for surgery. The image-based medical diagnosis expert system is crucial for a brain tumor patient. In this study, we combined two Magnetic Resonance Imaging (MRI)-based image datasets from Figshare and Kaggle to identify brain tumor MRI using a variety of convolutional neural network designs. To achieve competitive performance, we employ several data preprocessing techniques, such as resizing and enhancing contrast. The image augmentation techniques (E.g., rotated, width shifted, height shifted, shear shifted, and horizontally flipped) are used to increase data size, and five pre-trained models employed, including VGG-16, VGG-19, ResNet-50, Xception, and Inception-V3. The model with the highest accuracy, ResNet-50, performs at 96.76 percent. The model with the highest precision overall is Inception V3, with a precision score of 98.83 percent. ResNet-50 performs at 96.96% for F1-Score. The prominent accuracy of the implemented model, i.e., ResNet-50, compared with several earlier studies to validate the consequence of this introspection. The outcome of this study can be used in the medical diagnosis of brain tumors with an MRI-based expert system.

Large‐scale 3D fast Fourier transform computation on a GPU

AbstractWe propose a novel graphics processing unit (GPU) algorithm that can handle a large‐scale 3D fast Fourier transform (i.e., 3D‐FFT) problem whose data size is larger than the GPU's memory. A 1D FFT‐based 3D‐FFT computational approach is used to solve the limited device memory issue. Moreover, to reduce the communication overhead between the CPU and GPU, we propose a 3D data‐transposition method that converts the target 1D vector into a contiguous memory layout and improves data transfer efficiency. The transposed data are communicated between the host and device memories efficiently through the pinned buffer and multiple streams. We apply our method to various large‐scale benchmarks and compare its performance with the state‐of‐the‐art multicore CPU FFT library (i.e., fastest Fourier transform in the West [FFTW]) and a prior GPU‐based 3D‐FFT algorithm. Our method achieves a higher performance (up to 2.89 times) than FFTW; it yields more performance gaps as the data size increases. The performance of the prior GPU algorithm decreases considerably in massive‐scale problems, whereas our method's performance is stable.

Generalized extreme value analysis of criticality tallies in Monte Carlo calculation

Generalized extreme value (GEV) is the standard statistical model on the occurrences of extreme observations. The trinity theorem in GEV asserts that properly normalized maxima from blocks of data converge in distribution to one of the Weibull, Gumbel and Fréchet distributions as the data size increases. In this work, the methodology of GEV is applied to criticality tallies in Monte Carlo fission source cycles in order to evaluate the utility value of the distribution tail ends. Numerical results obtained under a sufficiently large number of particles per cycle show that the extreme value index (EVI) in GEV is not capable of being differentiated for the upper and lower distribution tail ends of criticality tallies. It is also shown that the EVI of criticality tallies falls within the range of Weibull distribution including the EVI of Gumbel distribution as the role of a boundary value layer. As a demonstration aimed at practicality, GEV is utilized for the population diagnosis under an insufficient number of particles per cycle. It turns out that the transition from one equilibrium to other equilibrium due to small population size makes the EVIs of upper and lower distribution tail ends depart from each other so that one of them falls in the range of Weibull distribution and the other in that of Fréchet distribution.

Antenna tracking profile generation using mixed-integer linear programming

The increasing demand for the improved quantitative and qualitative performance of Earth observation images requires an increase in image data sizes and satellite agility, and that requires an increased effectiveness of antenna operations transmitting data to a ground station. Here, a mixed-integer linear programming (MILP)-based algorithm generating a tracking profile for an antenna subjected to operational constraints was utilized to maximize data transmission time. The constraints included angle and angular velocity limits. This algorithm considers a transition between two sets of angles for the elevation-over-azimuth mechanism to avoid rapid movement. Data transmission angle ranges are depicted as polygons representing the mission environment as mathematical equations in the MILP model. To reduce computation time, two assumptions were applied to simplify the algorithm, which extracts sampling points that can represent interior points. The proposed algorithm was verified by general imaging mission scenarios from a high-mobility satellite with special mission scenarios (i.e., worst-case scenarios for antenna maneuvers). The results were then compared with the preliminary tracking profile from the ground station searching method, which is a heuristic method. The MILP-based algorithm created solutions where data could be transmitted for 97.87% of the entire contact duration; total maneuvers of azimuth and elevation angles averaged at below 45.49% and 60.88% than of those constantly directed at the ground station, respectively. Operational constraints were satisfied in all scenarios, confirming that the algorithm is well suited to both general and worst-case scenario missions.

Izindaba-Tindzaba: Machine learning news categorisation for Long and Short Text for isiZulu and Siswati

Local/Native South African languages are classified as low-resource languages. As such, it is essential to build the resources for these languages so that they can benefit from advances in the field of natural language processing. In this work, the focus was to create annotated news datasets for the isiZulu and Siswati native languages based on news topic classification tasks and present the findings from these baseline classification models. Due to the shortage of data for these native South African languages, the datasets that were created were augmented and oversampled to increase data size and overcome class classification imbalance. In total, four different classification models were used namely Logistic regression, Naive bayes, XGBoost and LSTM. These models were trained on three different word embeddings namely Bag-Of-Words, TFIDF and Word2vec. The results of this study showed that XGBoost, Logistic Regression and LSTM, trained from Word2vec performed better than the other combinations.

NySALT: Nyström-type inference-based schemes adaptive to large time-stepping

Large time-stepping is important for efficient long-time simulations of deterministic and stochastic Hamiltonian dynamical systems. Conventional structure-preserving integrators, while being successful for generic systems, have limited tolerance to time step size due to stability and accuracy constraints. We propose to use data to innovate classical integrators so that they can be adaptive to large time-stepping and are tailored to each specific system. In particular, we introduce NySALT, Nyström-type inference-based schemes adaptive to large time-stepping. The NySALT has optimal parameters for each time step learnt from data by minimizing the one-step prediction error. Thus, it is tailored for each time step size and the specific system to achieve optimal performance and tolerate large time-stepping in an adaptive fashion. We prove and numerically verify the convergence of the estimators as data size increases. Furthermore, analysis and numerical tests on the deterministic and stochastic Fermi-Pasta-Ulam (FPU) models show that NySALT enlarges the maximal admissible step size of linear stability, and quadruples the time step size of the Störmer–Verlet and the BAOAB when maintaining similar levels of accuracy.

Performance comparison of Dask and Apache Spark on HPC systems for neuroimaging

SummaryThe general increase in data size and data sharing motivates the adoption of Big Data strategies in several scientific disciplines. However, while several options are available, no particular guidelines exist for selecting a Big Data engine. In this paper, we compare the runtime performance of two popular Big Data engines with Python APIs, Apache Spark, and Dask, in processing neuroimaging pipelines. Our experiments use three synthetic neuroimaging applications to process the 606 GB BigBrain image and an actual pipeline to process data from thousands of anatomical images. We benchmark these applications on a dedicated HPC cluster running the Lustre file system while using varying combinations of the number of nodes, file size, and task duration. Our results show that although there are slight differences between Dask and Spark, the performance of the engines is comparable for data‐intensive applications. However, Spark requires more memory than Dask, which can lead to slower runtime depending on configuration and infrastructure. In general, the limiting factor was the data transfer time. While both engines are suitable for neuroimaging, more efforts need to be put to reduce the data transfer time and the memory footprint of applications.

Supervised learning techniques for dairy cattle body weight prediction from 3D digital images.

Introduction: The use of automation and sensor-based systems in livestock production allows monitoring of individual cows in real-time and provides the possibility of early warning systems to take necessary management actions against possible anomalies. Among the different RT monitoring parameters, body weight (BW) plays an important role in tracking the productivity and health status. Methods: In this study, various supervised learning techniques representing different families of methods in the machine learning space were implemented and compared for performance in the prediction of body weight from 3D image data in dairy cows. A total of 83,011 records of contour data from 3D images and body weight measurements taken from a total of 914 Danish Holstein and Jersey cows from 3 different herds were used for the predictions. Various metrics including Pearson's correlation coefficient (r), the root mean squared error (RMSE), and the mean absolute percentage error (MAPE) were used for robust evaluation of the various supervised techniques and to facilitate comparison with other studies. Prediction was undertaken separately within each breed and subsequently in a combined multi-breed dataset. Results and discussion: Despite differences in predictive performance across the different supervised learning techniques and datasets (breeds), our results indicate reasonable prediction accuracies with mean correlation coefficient (r) as high as 0.94 and MAPE and RMSE as low as 4.0 % and 33.0 (kg), respectively. In comparison to the within-breed analyses (Jersey, Holstein), prediction using the combined multi-breed data set resulted in higher predictive performance in terms of high correlation coefficient and low MAPE. Additional tests showed that the improvement in predictive performance is mainly due to increase in data size from combining data rather than the multi-breed nature of the combined data. Of the different supervised learning techniques implemented, the tree-based group of supervised learning techniques (Catboost, AdaBoost, random forest) resulted in the highest prediction performance in all the metrics used to evaluate technique performance. Reported prediction errors in our study (RMSE and MAPE) are one of the lowest in the literature for prediction of BW using image data in dairy cattle, highlighting the promising predictive value of contour data from 3D images for BW in dairy cows under commercial farm conditions.

Toward Enabling Performance-Guaranteed Networking in Next-Generation Cellular Networks

A blueprint for ultra-low latency in 5G cellular networks is designed to enable ultra-reliable low-latency communication services that require fast delivery of small data units (e.g., packets or frames). However, futuristic applications that are envisioned to be time-critical demand much more than what this blueprint can handle because their data units are typically very large. As the data size increases, the latency is greatly affected by how much data the network can transmit per unit time (i.e., bandwidth). This impact of bandwidth on latency brings the need to guarantee the bandwidth required to keep the latency within the desired time. In 5G, network slicing is introduced to guarantee performance, but how to handle the inherent nature of changing data unit size and radio channel quality over time remains an open question. In this article, for a detailed understanding, we first discuss end-to-end latency at the application level from the perspective of first-byte delay and transmission delay for the remaining bytes. We then investigate how recent techniques relate to end-to-end latency reduction, and present challenging issues caused by their inherent natures. To handle the issues, we propose a new design for next-generation (6G) cellular networks to provide performance guarantees, and discuss open issues in our network design.

CPV Tahmininde Makine Öğrenme Yöntemlerinin Performanslarının Değerlendirilmesi: Farklı Bölgeler ve Parametreler ile Büyük Veri Uygulaması

Scopa: In the project, machine learning methods that can interpret health data have been applied specifically to Canine parvovirus infection. While CPV can be diagnosed based on clinical findings, it needs to be supported by laboratory findings to distinguish it from other infections. Correct diagnosis is vital to distinguish CPV from other infections with bloody diarrhoea, which can result in death in puppies. For this reason, by analysing the virus together with other data that may be affected by the virus, the methods of making the most accurate decision were compared and evaluated. Purpose: In this study, it was aimed to interpret CPV, which is considered to be one of the most important infectious agents of dogs, popularly known as mad-head disease, using K-NearestNeighbour (KNN), RandomForest (RF), Logistic Regression and NaiveBayes classification algorithms in terms of different parameters. When the total accuracy values were examined, the accuracy rates decreased in logistic regression and RF methods when the insignificant variable was removed in the model. Result: RF method made the best predictions when Platelets, Platelet (PLT) variable was in the model.In cases where we do not want to remove this variable from the model, it can give us very efficient results. KNN method gives better results when the number of variables decreases. Especially when the data size increases, it has been observed that the machine learning method gives more efficient results with better performance. Keywords: Big Data, Data Mining, Big Data in Health, Prediction Algorithms, Canine parvovirus (CPV)

Modified recurrent equation-based cubic spline interpolation for missing data recovery in phasor measurement unit (PMU)

Background: Smart grid systems require high-quality phasor measurement unit (PMU) data for proper operation, control, and decision-making. Missing PMU data may lead to improper actions or even blackouts. While the conventional cubic interpolation methods based on the solution of a set of linear equations to solve for the cubic spline coefficients have been applied by many researchers for the interpolation of missing data, the computational complexity increases non-linearly with increasing data size. Methods: In this work, a modified recurrent equation-based cubic spline interpolation procedure for recovering missing PMU data is proposed. The recurrent equation-based method makes the computations of spline constants simpler. Using PMU data from the State Load Despatch Center (SLDC) in Madhya Pradesh, India, a comparison of the root mean square error (RMSE) values and time of calculation (ToC) is calculated for both methods. Results: The modified recurrent relation method could retrieve missing values 10 times faster when compared to the conventional cubic interpolation method based on the solution of a set of linear equations. The RMSE values have shown the proposed method is effective even for special cases of missing values (edges, continuous missing values). Conclusions: The proposed method can retrieve any number of missing values at any location using observed data with a minimal number of calculations.