Large Data Processing Research Articles

Reverse Inference: Decoding of Brain Activity and Cognitive Process

Cognitive neuroscientists rely on functional neuroimaging techniques and behavioral assays to investigate correlation between brain activation and cognitive processes. Researchers would also infer what cognitive process is engaged under a condition based on neuroimaging data. This is referred to as reverse inference or encoding paradigm and it has generated longstanding discussions among cognitive neuroscientists, data scientists and philosophers. The consensus is that it should be done with rigorous statistical methods and careful interpretations. Statistical methods and large collaborative databases and data processing tools have been developed to build classificatory or predictive models of reverse inference. However, these tools are not without pitfalls. The problem of cognitive process further complicates the problem since it is a latent variable and the wide discrepancy on the taxonomy and ontology of cognitive processes within the field. Online collaborative project has been set up to combat this problem, while others try to circumvent pre-established concepts by extracting latent variables from existing data or by focusing on the evolutionary prerequisite of cognition. A fundamental limitation of reverse inference is its dependency on correlational data, which lacks the explanatory power interventionist studies have. Models that seek to establish causal relation from time-series data can only provide weak inferences. Furthermore, the dynamic complexity of the brain network complicates the mechanism. This review provides an overview of reverse inference in cognitive neuroscience. It discusses advances in methods, limitations, and conceptual issues inherent within reverse inference, particularly addressing the challenges mentioned above.

OPTIMIZATION OF THE PARAMETERS OF SYNTHESIZED SIGNALS USING LINEAR APPROXIMATIONS BY THE NELDER-MEAD METHOD

Context. The article presents the results of a study of the effectiveness of using the Nelder-Mead method to optimize the parameters of linear approximations of synthesized signals. Algorithms have been developed and tested that integrate spectral, temporal, and statistical analyzes and provide reasonable optimization. The effectiveness of the application of the Nelder-Mead method was proven by experiment. The obtained results substantiate the improvement of the properties of the mutual correlation of signals and the reduction of the maximum deviations of the side lobes, which opens up prospects for the further application of the method in complex scenarios of signal synthesis. Objective. The purpose of the work is to evaluate the effectiveness of the application of the Nelder-Mead method when adjusting the parameters of linear approximations to optimize the mutual correlation and minimize side deviations of complex synthesized signals. Method. The main research method is the comparison of various optimization algorithms for the selection of the most effective approaches in linear approximations of synthesized signals, taking into account such criteria as accuracy, speed and minimization of deviations. Scientific works [1, 2, 4–6, 8, 9] present algorithms, including the Nelder-Mead method and differential evolution. The effectiveness of these methods is achieved due to adaptive optimization procedures that improve the characteristics of signals. It is worth noting that the methods have disadvantages associated with high requirements for computing resources, especially when processing large data. This can be minimized using combined optimization methods that take into account the interaction of signal parameters. Another important direction of improvement is the optimization of methods for adaptation to dynamic changes in the characteristics of complex signals, which allows to achieve high adaptability and reliability of real-time systems. Results. As a result of the experiment using the Nelder-Mead method, an increase in the similarity of spectral densities was achieved from 0.52 in the first iteration to 0.90 in the fourth, with a significant decrease in the distance between the peaks of the spectrum from 1.2 to 0.4, which indicates high adaptability and the accuracy of the method in adjusting the parameters of the synthesizedsignals. Conclusions. The effectiveness of the Nelder-Mead method for adjusting the specified parameters of the synthesized signals was experimentally proven, which is confirmed by a significant improvement in the similarity of the spectra with each iteration. This opens the way for additional optimizations and application of the method in various technological areas.

Effect of Architecture and Inference Parameters of Artificial Neural Network Models in the Detection Task on Energy Demand

Artificial neural network models for the task of detection are used in many fields and find various applications. Models of this kind require adequate computational resources and thus require adequate energy expenditure. The increase in the number of parameters, the complexity of architectures, and the need to process large data sets significantly increase energy consumption, which is becoming a key sustainability challenge. Optimization of computing and the development of energy-efficient hardware technologies are essential to reduce the energy footprint of these models. This article examines the effect of the type of model, as well as its parameters, on energy consumption during inference. For this purpose, sensors built into the graphics card were used, and software was developed to measure the energy demand of the graphics card for different architectures of YOLO models (v8, v9, v10), as well as for different batch and model sizes. This study showed that the increase in energy demand is not linearly dependent on batch size. After a certain level of batch size, the energy demand begins to decrease. This dependence does not occur only for n/t size models. Optimum utilization of computing power due to the number of processed images for the studied models occurs at the maximum studied batch size. In addition, tests were conducted on an embedded device.

Efficient equivalent-layer technique for processing irregularly spaced magnetic data on uneven surfaces

We present a new rapid and robust equivalent layer method for processing large magnetic data sets that can deal with irregularly spaced data points on undulating surfaces. Our method estimates an equivalent layer by numerically solving the upward continuation integral for a given set of magnetic data. While this estimated layer allows performing specific spatial transformations such as interpolation and continuations of any potential-field data, it cannot be used to perform conventional phase transformations on magnetic data, such as the reduction to the pole, because it requires a magnetic moment distribution over the equivalent layer of dipoles. To overcome this drawback, we transform the estimated equivalent layer into a planar equivalent layer of dipoles that equally reproduces the observed magnetic data. To carry out this transformation, we first use the fact that the estimated layer obtained by solving the upward continuation integral is a scaled version of the magnetic data on the layer plane, i.e., it can be considered a harmonic function. Then we use the upward continuation integral to continue the estimated layer into a planar and regular grid, which results in a second equivalent layer. Finally, we apply a fast 2D discrete convolution to transform this second layer into a planar equivalent layer of dipoles that also fits the observed magnetic data. Applications to synthetic and real data show that our methodology is able to efficiently process magnetic data, combining the robustness of accommodating irregularly spaced data with the ability to perform phase transformations.

Optimizing neural network models for predicting nuclear reactor channel temperature: A study on hyperparameter tuning and performance analysis

This study emphasizes how important accurate prediction of channel temperatures in nuclear reactors is for safety and operational efficiency. While traditional methods require long and complex processes such as kernel modeling and mathematical simulations, artificial neural networks (ANN) provide more efficient predictions by accelerating this process. The superior ability of ANNs to process large data sets is intended to demonstrate that this study will provide a valuable alternative compared to conventional methods and increase the accuracy of reactor temperature predictions. In this study, the training performances of Artificial Neural Network (ANN) developed to determine the nuclear reactor channel temperature with different hyperparameter combinations were analysed. It was conducted several experimental studies to assess the influence of hyperparameters on our model for nuclear reactor parameter data prediction. The training and validation results indicates that learning rate, hidden layer sizes and number have critical effects for the more precisive prediction. It was observed that models with a learning rate of 0.05 and 0.5 achieved successful learning with less fluctuation in training and validation errors. When looking at hidden layer sizes, networks with 32 and 64 neurons performed better than networks with 16 neurons. For the test phase our model can successfully predict data with slight error margin. As a result, we demonstrated that neural networks are a powerful tool in nuclear reactor channel temperature prediction through our proposed model.

Direct energy deposition for smart micro reactor

ABSTRACT Nuclear microreactors (MRs) have attracted significant attention for their versatility in installation locations, especially with the growing demand for sustainable green energy in large data processing facilities. However, ensuring their safe operation necessitates frequent inspections, which can be challenging for practical efficient management. This study proposes a novel approach using direct energy deposition to incorporate an optical fiber sensor into MR components, enabling real-time monitoring with artificial intelligence. The embedded optical fiber generates real-time data that allows for AI-driven in-vivo thermal deformation analysis. Our smart MR component can detect structural anomalies, identify abnormal operations, and assess operational conditions through augmented reality interfaces and AI technology.

Интеллектуальная обработка текстовой информации: обзор автоматизированных методов суммаризации

Interest in innovative technological strategies and modern digital tools has increased significantly due to the need to manage large amounts of unstructured data. This paper reviews current paradigms and services for automated summarization, developed based on interdisciplinary research in linguistics, computer technologies, and artificial intelligence. It focuses on syntactic and lexical techniques employed by neural network models for text compression. The paper presents performance examples of such AI-powered services as QuillBot, Summate.it, WordTune, SciSummary, Scholarcy, and OpenAI ChatGPT. The contemporary automated models proved effective in using extractive and abstractive methods to generate summaries of varying quality and length. The extractive approach relies on identifying the most significant sentences from the original text, while abstractive algorithms create new sentence structures that preserve the main idea of the original content. Automated summarizers effectively utilize text compression techniques that are inherent to human approach to text processing, e.g., they exclude redundant information, simplify complex structures, and generalize data. These technologies provide high accuracy and coherence in the generated summaries, though each summarization model has its limitations. Optimal results depend on the specifics of the task at hand: extractive models provide brevity and precision while abstractive ones allow for deeper semantic processing. Automated summarization is becoming an important tool in various fields that require effective analysis and processing of large text data.

Improved multi-focus image fusion using online convolutional sparse coding based on sample-dependent dictionary

Multi-focus image fusion merges multiple images captured from different focused regions of a scene to create a fully-focused image. Convolutional sparse coding (CSC) methods are commonly employed for accurate extraction of focused regions, but they often disregard computational costs. To overcome this, an online convolutional sparse coding (OCSC) technique was introduced, but its performance is still limited by the number of filters used, affecting overall performance negatively. To address these limitations, a novel approach called Sample-Dependent Dictionary-based Online Convolutional Sparse Coding (SCSC) was proposed. SCSC enables the utilization of additional filters while maintaining low time and space complexity for processing high-dimensional or large data. Leveraging the computational efficiency and effective global feature extraction of SCSC, we propose a novel method for multi-focus image fusion. Our method involves a two-layer decomposition of each source image, yielding a base layer capturing the predominant features and a detail layer containing finer details. The amalgamation of the fused base and detail layers culminates in the reconstruction of the final image. The proposed method significantly mitigates artifacts, preserves fine details at the focus boundary, and demonstrates notable enhancements in both visual quality and objective evaluation of multi-focus image fusion.

DONGLE 2.0: Direct FPGA-Orchestrated NVMe Storage for HLS

Rapid growth in data size poses significant computational and memory challenges to data processing. FPGA accelerators and near-storage processing have emerged as compelling solutions for tackling the growing computational and memory requirements. Many FPGA-based accelerators have shown to be effective in processing large data sets by leveraging the storage capability of either host-attached or FPGA-attached storage devices. However, the current HLS development environment does not allow direct access to host-or FPGA-attached NVMe storage from the HLS code. As such, users must frequently hand off between HLS and host code to access data in storage, and such a process requires tedious programming to ensure functional correctness. Moreover, since the HLS code uses radically different methods to access storage compared to DRAM, the HLS codebase targeting DRAM-based platforms cannot be easily ported to NVMe-based platforms, resulting in limited code portability and reusability. Furthermore, frequent suspension of HLS kernel and synchronization between CPU and FPGA introduce significant latency overhead and require sophisticated scheduling mechanisms to hide latency. To address these challenges, we propose a new HLS storage interface named DONGLE 2.0 that enables direct FPGA-orchestrated NVMe storage access. By providing a unified interface for storage and memory access, DONGLE 2.0 allows a single-source HLS program to target multiple memory/storage devices, thus making the codebase cleaner, portable, and more efficient. DONGLE 2.0 is an extension to DONGLE 1.0 [ 1 ] but adds support for host-attached storage. While its primary focus is still on FPGA NVMe access in near-storage configurations, the added host storage support ensures its compatibility with platforms that lack native support for FPGA-attached NVMe storage. We implemented a prototype of DONGLE 2.0 using an AMD/Xilinx Alveo U200 FPGA and Solidigm DC-P4610 SSD. Our evaluation on various workloads showed a geometric mean speed-up of 2.3× and a reduction in lines of code (LoC) by 2.4× compared to the state-of-the-art commercial platform when using FPGA-attached NVMe storage. Moreover, DONGLE 2.0 demonstrated a geometric mean speed-up of 1.5× and a reduction in LoC by 2.4× compared to the state-of-the-art commercial platform when using host-attached NVMe storage.

DASEventNet: AI‐Based Microseismic Detection on Distributed Acoustic Sensing Data From the Utah FORGE Well 16A (78)‐32 Hydraulic Stimulation

AbstractDistributed acoustic sensing (DAS) has emerged as a promising seismic technology for monitoring microearthquakes (MEQs) with high spatial resolution. Efficient algorithms are needed for processing large DAS data volumes. This study introduces a deep learning (DL) model based on a Residual Convolutional Neural Network (ResNet) for detecting MEQs using DAS data, named as DASEventNet. The test data were collected from the Utah FORGE 16A (78)‐32 hydraulic stimulation experiments conducted in April 2022. The DASEventNet model achieves a remarkable accuracy of 100% when discriminating MEQs from noise in the raw test set of 260 examples. Surprisingly, the model identified weak MEQ signatures that have been manually categorized as noise. The decision‐making process with the model is decoded by the classic activation map, which illuminates learning features of the DASEventNet model. These features provide clear illustrations of weak MEQs and varied noise types. Finally, we apply the trained model to the entire period (∼7 days) of continuous DAS recordings and find that it discovers >5,700 new MEQs, previously unregistered in the public Silixa DAS catalog. The DASEventNet model significantly outperforms the traditional seismic method Short‐Term Average/Long‐Term Average (STA/LTA), which detected only 1,307 MEQs. The DASEventNet detection threshold is Mw−1.80 compared to the minimum magnitude of Mw−1.14 detected by STA/LTA. The spatiotemporal distribution of the newly identified MEQs defines an extensive stimulation zone and more accurately characterizes fracture geometry. Our results highlight the potential of DL for long‐term, real‐time microseismic monitoring that can improve enhanced geothermal systems and other activities that include subsurface hydraulic fracturing.

A New Hybrid Model for Heart Disease Prediction Using Machine Learning Algorithms Optimized by Modified Whale Optimization Algorithm

Heart disease is a kind of cardiovascular disease (CVD) which globally considered death's number one cause. Data Science plays an important role in processing large data amounts in the healthcare domain. There are several problems that could hinder appropriate cardiac monitoring, including limited of medical dataset, lack of depth analysis, and feature selection. In this paper, we exploit the method of Fast Correlation-Based Feature Selection (FCBF) for filtering extra features for developing heart disease classification quality. Next, we implement the classification according to various algorithms of classification like Decision tree, Naïve Bayes, Logistic Regression, K-Nearest Neighbour (KNN), Random Forest, Support Vector Machine (SVM), and a Multilayer Perception which is optimized with the modified whale optimization algorithm evaluation version named Modified Whale Optimization Algorithm (MWOA). The proposed model optimized with FCBF as well as MWOA obtain an 83.26% accuracy score by Logistic Regression.

Large Language Models Can Enable Inductive Thematic Analysis of a Social Media Corpus in a Single Prompt: Human Validation Study.

Manually analyzing public health-related content from social media provides valuable insights into the beliefs, attitudes, and behaviors of individuals, shedding light on trends and patterns that can inform public understanding, policy decisions, targeted interventions, and communication strategies. Unfortunately, the time and effort needed from well-trained human subject matter experts makes extensive manual social media listening unfeasible. Generative large language models (LLMs) can potentially summarize and interpret large amounts of text, but it is unclear to what extent LLMs can glean subtle health-related meanings in large sets of social media posts and reasonably report health-related themes. We aimed to assess the feasibility of using LLMs for topic model selection or inductive thematic analysis of large contents of social media posts by attempting to answer the following question: Can LLMs conduct topic model selection and inductive thematic analysis as effectively as humans did in a prior manual study, or at least reasonably, as judged by subject matter experts? We asked the same research question and used the same set of social media content for both the LLM selection of relevant topics and the LLM analysis of themes as was conducted manually in a published study about vaccine rhetoric. We used the results from that study as background for this LLM experiment by comparing the results from the prior manual human analyses with the analyses from 3 LLMs: GPT4-32K, Claude-instant-100K, and Claude-2-100K. We also assessed if multiple LLMs had equivalent ability and assessed the consistency of repeated analysis from each LLM. The LLMs generally gave high rankings to the topics chosen previously by humans as most relevant. We reject a null hypothesis (P<.001, overall comparison) and conclude that these LLMs are more likely to include the human-rated top 5 content areas in their top rankings than would occur by chance. Regarding theme identification, LLMs identified several themes similar to those identified by humans, with very low hallucination rates. Variability occurred between LLMs and between test runs of an individual LLM. Despite not consistently matching the human-generated themes, subject matter experts found themes generated by the LLMs were still reasonable and relevant. LLMs can effectively and efficiently process large social media-based health-related data sets. LLMs can extract themes from such data that human subject matter experts deem reasonable. However, we were unable to show that the LLMs we tested can replicate the depth of analysis from human subject matter experts by consistently extracting the same themes from the same data. There is vast potential, once better validated, for automated LLM-based real-time social listening for common and rare health conditions, informing public health understanding of the public's interests and concerns and determining the public's ideas to address them.

Computer-assisted regional-residual gravity anomaly separation with regularized first- and second-order derivatives

The regional-residual separation of gravity anomalies in crustal and mineral exploration was a graphical-based procedure before the advent of fast digital computers and the need for more efficient algorithms to process large data sets. However, since requiring the supervision of an experienced interpreter, the results, once obtained with graphical procedures, are often accepted as second to none in producing anomalies with geologic significance. Numerical methods based on spectral filtering and robust polynomial fitting have worked in many scenarios but seem not fully effective in replicating in algorithms the kind of results once obtained with interpreter-assisted graphical methods. We develop a computer-assisted regional residual separation (CARRS) procedure implemented by a set of short MATLAB scripts; in many aspects, CARRS simulates the operations of former graphical methods but requires few decisions and minor handwork from the interpreter. CARRS applies robust polynomial fitting to points with a low horizontal gradient and a vertical second-order derivative; thus, selecting fitting points as a real interpreter would do with graphical approaches in outlining the regional field. A regularized procedure is used to calculate the stable first- and second-order derivatives. Companion MATLAB codes allow for the replication of our results and further exploration with different threshold levels to identify flat domain regions. CARRS is illustrated with airborne gravity data CPRM-1123, freely available to download. A data window of the residual field is used to analyze the distribution of cassiterite deposits in greisen zones of the Paleoproterozoic Velho Guilherme granites in the Amazon Craton, as their distribution appears in the observed gravity anomaly and its corresponding residual field.

Revitalizing the Forensic Accounting: An Exploratory Study on Mitigating the Financial Risk using Data Analytics

Risk mitigation and fraud prevention in the present times has more focus on digital metadata, wherein forensic accountants are required to make use of robust IT techniques and tools. Data Analytics has got huge implications for forensic accounting. With the ever-increasing electronic element of frauds in the present age of digitalization, and complexity of financial transactions and instruments, the challenges are growing multifold for the forensic accounting profession. The use of data analytics techniques, which enables the processing of large data in almost no time, simplify the task of forensic accountant to a large extent. It enables the identification of patterns and anomalies with the use of machine learning, deep learning, natural language processing, data mining and similar other statistical techniques. Beneish M-score and the Altman Z-score models are proven effective forensic accounting tools in fraud detection and insolvency prediction. This study has been undertaken to test the effectiveness and efficiency of these forensic tools in fraud detection and insolvency prediction with the case study of Bhushan Steel Ltd. Based upon the findings of the study, it is suggested that these models should be integrated with the data analytics, as these are highly effective and efficient in detection of financial statements frauds and this integration would equip the naïve investor in selection of financially sound company(s) to invest in. Data analytics is thus indispensable in the field of forensic accounting in the present era. Since, there is increased funding by many international funding agencies including the World Bank for empowering the forensic accountants and increasing awareness towards this profession, especially in developing countries; it can be hoped that the integration of data analytics with forensic accounting would gain momentum soon.

Predictive Analysts Using Different Algorithm Machine And Deep Learning For Financial Market: Review Paper

This paper presents the development and effectiveness of deep and machine learning techniques in forecasting stock market trends. This paper review focus on key developments from 2020 to 2024 involves the integration of ensemble learning, sentiment analysis from social media, and various predictive algorithms such as LSTM, CNN, and ANN. These techniques improves predication accuracy by efficient processing large data groups and detect nonlinear patterns in stock price trends. in addition, the paper presented background about various machine learning algorithms such as KNN, naive Bayes, linearregression decision trees, SVM, randomforests and deep learning models suchas neural network,LSTM and CNN .the findings of this study to providing valuable insights to researchers and practitioners who aim to improve investment strategies and improve predictive accuracy By looking at the appropriate algorithm

A Software Tool For Automated Analysis And Characterization Of Raw PIV Images

The time to process large PIV data sets can be expensive and the resultant velocity fields are a strong function of the quality of the particle images used. For new and even experienced users of PIV, determination of the quality of a particle image data set can be challenging and time consuming especially for a large number of data sets. This is compounded with new high-speed camera systems that are capable of collecting terabytes of data quickly. A software tool is described here that allows the user to make informed decisions on the general quality of the data sets and what pre-processing image data steps are needed. The software provides statistical feedback on such parameters as particle count, particle size, particle intensity for not only a single image but also a complete data set. Using this software allows the user to make informed decisions and generate and document the quality of the data collected. Based on this, a robust PIV processing algorithm can be developed.

TYMELINE WHITE PAPER - EXPLORING TEAM PRODUCTIVITY THROUGH THE APPLICATION OF ARTIFICIAL INTELLIGENCE AND BLOCKCHAIN TECHNOLOGY

Tymeline is an AI-driven platform that revolutionizes goal setting and performance management by leveraging historical data securely integrated via blockchain. This platform enhances productivity across various industries through adaptive, data-driven insights. It addresses the limitations of traditional project management and CRM tools, which often fail to effectively utilize historical performance data. Tymeline’s proprietary AI processes large data volumes to identify trends, predict outcomes, and generate personalized goal roadmaps. This paper details Tymeline’s innovative approach, underlying technologies, key features, and substantial benefits, showcasing its potential to transform team and organizational productivity.

Artificial Intelligence in Health Professions Education

Artificial Intelligence (AI) is revolutionizing various fields, including Health Professions Education (HPE), with its ability to mimic human problem-solving and decision-making capabilities. Unlike the self-aware AI depicted in movies, current AI systems, like ChatGPT, follow human commands and have shown remarkable growth, reaching one million users within five days. AI’s strengths include processing large data sets rapidly, personalizing learning, automating tasks, enhancing presentations, generating clinical cases, and providing real-time feedback.

ПРОГНОЗИРОВАНИЕ ВЫРУЧКИ БЫСТРОРАСТУЩЕЙ КОМПАНИИ С ИСПОЛЬЗОВАНИЕМ ЛОГИСТИЧЕСКОЙ КРИВОЙ

The aim of the study is to assess the advancement potential of fast-growing companies, to create prerequisites for predicting the gain of such companies. The objective is to identify and forecast the growth points of the country’s economy, which is generally based on the advancement of individual economic entities, namely organizations. The research method for forecasting revenue as the main indicator of fast-growing companies (FGC) is considered. As an approximation, a logistic curve (the Ferhulst curve) is examined, the parameters of which are identified by the least squares method. Open data from financial statements of organizations are applied as a source. This work uses the criterion for distinguishing fast-growing companies; the average annual growth is of at least 50% at current prices. The novelty of the work lies in studying fast-growing organizations using a logistic curve (sigmoid). Approximation parameters are identified. A growth assessment (forecast) for organizations for 2040 is made. The statistical reliability of the selected approximation is established for 714 large and medium-sized fast-growing companies from the sample. The research results state, that a sample of about 900 fast-growing Russian companies is identified using the methods for processing large data sets from 2.5 million organizations. The illustrative calculations are based on actual sample data, highlighting the example of Wildberries in more detail. An assessment of the growth opportunities of the organizations under consideration is made. Studying the applicability of the proposed method using statistical criteria is carried out. The findings show that studying fast-growing organizations’ advancement for a medium or longer period using a logistic curve looks more preferable than applying an exponential one. For the vast majority of such organizations, the approximation of the logistic curve is statistically significant.

Average nucleotide identity-based Staphylococcus aureus strain grouping allows identification of strain-specific genes in the pangenome.

We analyzed the genomic diversity of Staphylococcus aureus, a globally prevalent bacterial species that causes serious infections in humans. Our goal was to build a genetic picture of the different strains of S. aureus and which genes may be associated with them. We reprocessed >84,000 genomes and subsampled to remove redundancy. We found that individual samples sharing >99.5% of their genome could be grouped into strains. We also showed that a portion of genes that are present in intermediate frequency in the species are strongly associated with some strains but completely absent from others, suggesting a role in strain specificity. This work lays the foundation for understanding individual gene histories of the S. aureus species and also outlines strategies for processing large bacterial genomic data sets.