Data Size Research Articles

Transfer learning enables identification of multiple types of RNA modifications using nanopore direct RNA sequencing

Nanopore direct RNA sequencing (DRS) has emerged as a powerful tool for RNA modification identification. However, concurrently detecting multiple types of modifications in a single DRS sample remains a challenge. Here, we develop TandemMod, a transferable deep learning framework capable of detecting multiple types of RNA modifications in single DRS data. To train high-performance TandemMod models, we generate in vitro epitranscriptome datasets from cDNA libraries, containing thousands of transcripts labeled with various types of RNA modifications. We validate the performance of TandemMod on both in vitro transcripts and in vivo human cell lines, confirming its high accuracy for profiling m6A and m5C modification sites. Furthermore, we perform transfer learning for identifying other modifications such as m7G, Ψ, and inosine, significantly reducing training data size and running time without compromising performance. Finally, we apply TandemMod to identify 3 types of RNA modifications in rice grown in different environments, demonstrating its applicability across species and conditions. In summary, we provide a resource with ground-truth labels that can serve as benchmark datasets for nanopore-based modification identification methods, and TandemMod for identifying diverse RNA modifications using a single DRS sample.

On nonlinear effects in multiphase WKB analysis for the nonlinear Schrödinger equation *

We consider the Schrödinger equation with an external potential and a cubic nonlinearity, in the semiclassical limit. The initial data are sums of WKB states, with smooth phases and smooth, compactly supported initial amplitudes, with disjoint supports. We show that like in the linear case, a superposition principle holds on some time interval independent of the semiclassical parameter, in several régimes in term of the size of initial data with respect to the semiclassical parameter. When nonlinear effects are strong in terms of the semiclassical parameter, we invoke properties of compressible Euler equations. For weaker nonlinear effects, we show that there may be no nonlinear interferences on some time interval independent of the semiclassical parameter, and interferences for later time, thanks to explicit computations available for particular phases.

A study of universal morphological analysis using morpheme-based, holistic, and neural approaches under various data size conditions

A study of universal morphological analysis using morpheme-based, holistic, and neural approaches under various data size conditions

High-speed light-sheet fluorescent microscopy with an adaptive three-dimensional region-of-signal identification method

Light-sheet fluorescent microscopy (LSFM) is a crucial 3D imaging technique for tissues with advantages in high 3D spatial resolution, effective optical sectioning, and quick imaging. Prior to imaging, the sample's imaging depth, imaging region, and region of interest (ROI) division should all be established in accordance with the sample's outermost dimensions. However, because biological samples are inherently irregular a traditional imaging flow that images ROI by ROI, slice by slice, would waste a great deal of time and data storage space by additionally recording non-signal regions. In order to solve this problem, we create a high-speed LSFM with an adaptive three-dimensional region-of-signal identification method (RSI). We immediately perform crude imaging for each ROI before the software analyzes the longitudinal distribution and, ultimately, uses dimensionality reduction, binarization, filtering, and redundancy expansion to acquire the signal region result. By imaging two mouse spinal cord segments with a total of 170 ROIs and 180 ROIs on two channels, respectively, the high-speed imaging capability is validated. Our method reduced the overall imaging time and raw data size by more than treble and 14 times, respectively, without losing signal. With a programmable imaging range based on the real signal region, this method offers an efficient tool for flexible light-sheet microscopy. It can also be applied to other microscopy techniques to speed up imaging.

Optimization of parallel SVM algorithm for big data

Parallel Support Vector Machine (SVM) based on big data has achieved some results in data mining, but due to the complexity of the data itself and a large amount of noisy data, its execution efficiency and classification accuracy in the big data environment are very low. In order to eliminate noise, a noise reduction method based on Noise Cleaning (NC) strategy was proposed, and redundant training samples in big data environments were deleted; Introduce an improved Artificial Fish Swarm Algorithm (IAFSA) to obtain the final Parallel SVM algorithm using mutual information and artificial fish swarm algorithm based on MapReduce (MIAFSA-PSVM) classification model. The results indicate that when compared to CMI-PSVM, the execution time of MIAFSA-PSVM algorithm is higher on the NDC dataset with the largest data size, The SVM parameter optimization algorithm based on MapReduce and cuckoo search (CSSVM-MR) and the particle swarm optimization based parallel support vector machine ensemble algorithm (PSO-PSVM) decreased by 40.1%, 79.3%, and 51.7%, respectively. This indicates that GIESVM-MR and MIAFSA-PSVM have strong adaptability to big data environments and high classification accuracy.

Dynamic Sensor Scheduling for Data Size Reduction in a Sensor Cloud System Based on Minimum Reconstruction Error

Dynamic Sensor Scheduling for Data Size Reduction in a Sensor Cloud System Based on Minimum Reconstruction Error

Georeferencing of Satellite Images with Geocoded Image Features

Abstract. Currently, using digital orthophoto map (DOM) and digital elevation model (DEM) as reference to achieve geometric positioning of newly acquired satellite images has become a popular photogrammetric approach. However, this method relies on DOM and DEM data which requires a lot of storage space in practical applications. In addition, for geometric positioning of satellite images, only sparse image feature points are needed as control points. Consequently, for the sake of convenience, the compression of control data emerges as a necessity with significant practical implications. This paper investigates a "cloud control" photogrammetry method based on geocoded image features. The method extracts SIFT feature points from DOMs, and obtains their ground coordinates, then constructs geocoded image feature library instead of DOM and DEM data as control, thus realizing the compression of control data. Experiments conducted on the Tianhui-1, Ziyuan-3 and Gaofen-2 satellite images demonstrate that the proposed method can achieve high-precision geometric positioning of satellite images and greatly reduce the size of the control data. Specifically, with the reduction of the reference data from 180~1248 MB 2 m DOM and 30 m DEM to 5~10 MB geocoded image features, the geopositional accuracies of the test Tianhui-1, Ziyuan-3 and Gaofen-2 images are improved from 3.12 pixels to 1.74 pixels, 3.69 pixels to 1.09 pixels, and 150.93 pixels to 2.67 pixels, respectively.

Transferability of Machine Learning Models for Geogenic Contaminated Groundwaters.

Machine learning models show promise in identifying geogenic contaminated groundwaters. Modeling in regions with no or limited samples is challenging due to the need for large training sets. One potential solution is transferring existing models to such regions. This study explores the transferability of high fluoride groundwater models between basins in the Shanxi Rift System, considering six factors, including modeling methods, predictor types, data size, sample/predictor ratio (SPR), predictor range, and data informing. Results show that transferability is achieved only when model predictors are based on hydrochemical parameters rather than surface parameters. Data informing, i.e., adding samples from challenging regions to the training set, further enhances the transferability. Stepwise regression shows that hydrochemical predictors and data informing significantly improve transferability, while data size, SPR, and predictor range have no significant effects. Additionally, despite their stronger nonlinear capabilities, random forests and artificial neural networks do not necessarily surpass logistic regression in transferability. Lastly, we utilize the t-SNE algorithm to generate low-dimensional representations of data from different basins and compare these representations to elucidate the critical role of predictor types in transferability.

Int&in: A machine learning-based web server for active split site identification in inteins.

Inteins are proteins that excise themselves out of host proteins and ligate the flanking polypeptides in an auto-catalytic process called protein splicing. In nature, inteins are either contiguous or split. In the case of split inteins, the two fragments must first form a complex for the splicing to occur. Contiguous inteins have previously been artificially split in two fragments because split inteins allow for distinct applications than contiguous ones. Even naturally split inteins have been split at unnatural split sites to obtain fragments with reduced affinity for one another, which are useful to create conditional inteins or to study protein-protein interactions. So far, split sites in inteins have been heuristically identified. We developed Int&in, a web server freely available for academic research (https://intein.biologie.uni-freiburg.de) that runs a machine learning model using logistic regression to predict active and inactive split sites in inteins with high accuracy. The model was trained on a dataset of 126 split sites generated using the gp41-1, Npu DnaE and CL inteins and validated using 97 split sites extracted from the literature. Despite the limited data size, the model, which uses various protein structural features, as well as sequence conservation information, achieves an accuracy of 0.79 and 0.78 for the training and testing sets, respectively. We envision Int&in will facilitate the engineering of novel split inteins for applications in synthetic and cell biology.

Deriving Input Variables through Applied Machine Learning for Short-Term Electric Load Forecasting in Eskilstuna, Sweden

As the demand for electricity, electrification, and renewable energy rises, accurate forecasting and flexible energy management become imperative. Distribution network operators face capacity limits set by regional grids, risking economic penalties if exceeded. This study examined data-driven approaches of load forecasting to address these challenges on a city scale through a use case study of Eskilstuna, Sweden. Multiple Linear Regression was used to model electric load data, identifying key calendar and meteorological variables through a rolling origin validation process, using three years of historical data. Despite its low cost, Multiple Linear Regression outperforms the more expensive non-linear Light Gradient Boosting Machine, and both outperform the “weekly Naïve” benchmark with a relative Root Mean Square Errors of 32–34% and 39–40%, respectively. Best-practice hyperparameter settings were derived, and they emphasize frequent re-training, maximizing the training data size, and setting a lag size larger than or equal to the forecast horizon for improved accuracy. Combining both models into an ensemble could the enhance accuracy. This paper demonstrates that robust load forecasts can be achieved by leveraging domain knowledge and statistical analysis, utilizing readily available machine learning libraries. The methodology for achieving this is presented within the paper. These models have the potential for economic optimization and load-shifting strategies, offering valuable insights into sustainable energy management.

A Novel Approach of Cloud Computing Network for Authentication and Security Enhancement of IoT Enabled Cancer Forecasting System

Recently, a variety of approaches have been employed to address a broad spectrum of real-world issues; these methodology cover a variety of areas, including healthcare systems. Previous researchers concentrated on health-care monitoring software. It had several shortcomings, such as poor health-care data storage, time, expense, and processing complexity. This paper proposes a unique IoT-enabled and secured clinical monitoring paradigm to address these issues. Initially, implant several sensors to gather information on vital indicators like body temperature fluctuation. Phone numbers, marital status, heart rate deviation, residence, name, age, and blood pressure are among the patient's health information. The IoT medical sensor dataset is used in this investigation. During pre-processing, extra unnecessary attributes are removed, resulting in data size reduction and normalization. A convolutional neural network supporting the classification of cancer sickness that is based on the improved teaching-learning optimization (CNN-ITLO) method. The CNN-ITLO model ascertains whether or not the patient is cancer-prone based on the sensor input. The management of the hospital receives the gathered data after which it is analyzed. Lastly, the homomorphic encryption approach is used to encrypt and store the patient's data on the cloud. The Java platform will be used to carry out the recommended work. According to the experiment results, the suggested method performed better than current cutting-edge practices.

Transition role of entangled data in quantum machine learning

Entanglement serves as the resource to empower quantum computing. Recent progress has highlighted its positive impact on learning quantum dynamics, wherein the integration of entanglement into quantum operations or measurements of quantum machine learning (QML) models leads to substantial reductions in training data size, surpassing a specified prediction error threshold. However, an analytical understanding of how the entanglement degree in data affects model performance remains elusive. In this study, we address this knowledge gap by establishing a quantum no-free-lunch (NFL) theorem for learning quantum dynamics using entangled data. Contrary to previous findings, we prove that the impact of entangled data on prediction error exhibits a dual effect, depending on the number of permitted measurements. With a sufficient number of measurements, increasing the entanglement of training data consistently reduces the prediction error or decreases the required size of the training data to achieve the same prediction error. Conversely, when few measurements are allowed, employing highly entangled data could lead to an increased prediction error. The achieved results provide critical guidance for designing advanced QML protocols, especially for those tailored for execution on early-stage quantum computers with limited access to quantum resources.

Enabling low-latency IoT communication for resource-constrained devices with the led cipher and decipher protocol

Block cipher algorithms are crucial for securing applications on resource-constrained devices. This paper introduces the modified light encryption device (MLED) cipher-decipher architecture, specifically designed to accommodate both 64-bit and 128-bit key sizes while maintaining a consistent 64-bit block and data size. MLED comprises 8-step and 12-step processes for MLED-64 and MLED-128 modules, respectively. Each stage involves a four-round operation followed by an add-round key operation. The add constant module (ACM) and mixed column modules (MCMs) within the round operation have been optimized for improved latency and throughput. Performance analysis reveals that MLED-64/128 requires less than 1% of the available slices and operates at 125 MHz on the Artix-7 FPGA. It achieves delays of 7.5 and 12.5 clock cycles for MLED-64 and MLED-128, respectively, translating to throughputs of 1366.5 Mbps and 819.89 Mbps. Additionally, MLED-64/128 exhibits hardware efficiencies of 2.373 and 0.986 Mbps/slice, respectively. Comparative evaluations with existing LED and other block ciphers (BCs) demonstrate that MLED-64/128 achieves significant improvements in latency, throughput, and efficiency, making it a compelling choice for securing resource-constrained IoT applications.

Distributed statistical learning algorithm for nonlinear regression with autoregressive errors

The growing size of modern data brings challenges to statistical learning, and substantial distributed algorithms have been proposed. However, most of them need the homogeneity assumption that the distribution of the local data is the same as that of the global data. This is seldom in practice, and the learning performance deteriorates seriously if this assumption is not satisfied. Moreover, they are only for independent data, and can not incorporate the serial correlations between data. To solve these issues, we propose a novel distributed statistical learning framework for the nonlinear regression with autoregressive errors, which realizes communication-efficient distributed optimization, and overcomes the homogeneity assumption. The theoretical results also guarantee that the new distributed framework is equivalent to the global one. Numerical experiments also illustrate the good performance of the new method.

Utilizing Convolutional Neural Networks (CNN) and U-Net Architecture for Precise Crop and Weed Segmentation in Agricultural Imagery: A Deep Learning Approach

This study presents the implementation and evaluation of a convolutional neural network (CNN) based image segmentation model using the U-Net architecture for forest image segmentation. The proposed algorithm starts by preprocessing the datasets of satellite images and corresponding masks from a repository source. Data preprocessing involves resizing, normalizing, and splitting the images and masks into training and testing datasets. The U-Net model architecture, comprising encoder and decoder parts with skip connections, is defined and compiled with binary cross-entropy loss and Adam optimizer. Training includes early stopping and checkpoint saving mechanisms to prevent overfitting and retain the best model weights. Evaluation metrics such as Intersection over Union (IoU), Dice coefficient, pixel accuracy, precision, recall, specificity, and F1-score are computed to assess the model's performance. Visualization of results includes comparing predicted segmentation masks with ground truth masks for qualitative analysis. The study emphasizes the importance of training data size in achieving accurate segmentation models and highlights the potential of U-Net architecture for forest image segmentation tasks.

Probabilistic neural networks for incremental learning over time-varying streaming data with application to air pollution monitoring

This paper proposes a novel algorithm for incremental learning over streaming data in a non-stationary environment. The idea refers to the applicability of Probabilistic Neural Networks (PNNs), commonly used as a fast and robust method for solving classification problems in a stationary scenario. It is well known that PNNs have solid mathematical foundations but they fail in the non-stationary scenario and are not able to deal with concept drift. In this paper, the method based on the orthogonal series expansions of unknown probability densities is proposed. This approach significantly extends the applicability of the PNNs. The nonparametric procedures called Incremental Probabilistic Neural Networks (IPNNs) for tracking drifting probability densities or conditional class densities, in the case of classification, are presented. We prove their convergence as the stream data size is growing to infinity. More precisely, we show convergence in probability and with probability one as the sample size tends to infinity. Compared to the existing literature, almost exclusively based on various heuristics, the proposed method is mathematically justified and has solid theoretical foundations. Simulations performed on synthetic and air pollution data confirm the effectiveness of the algorithm.

ELIMS: Ensemble Learning-Based Spatial Segmentation of Mass Spectrometry Imaging to Explore Metabolic Heterogeneity.

Spatial segmentation is an essential processing method for image analysis aiming to identify the characteristic suborgans or microregions from mass spectrometry imaging (MSI) data, which is critical for understanding the spatial heterogeneity of biological information and function and the underlying molecular signatures. Due to the intrinsic characteristics of MSI data including spectral nonlinearity, high-dimensionality, and large data size, the common segmentation methods lack the capability for capturing the accurate microregions associated with biological functions. Here we proposed an ensemble learning-based spatial segmentation strategy, named eLIMS, that combines a randomized unified manifold approximation and projection (r-UMAP) dimensionality reduction module for extracting significant features and an ensemble pixel clustering module for aggregating the clustering maps from r-UMAP. Three MSI datasets are used to evaluate the performance of eLIMS, including mouse fetus, human adenocarcinoma, and mouse brain. Experimental results demonstrate that the proposed method has potential in partitioning the heterogeneous tissues into several subregions associated with anatomical structure, i.e., the suborgans of the brain region in mouse fetus data are identified as dorsal pallium, midbrain, and brainstem. Furthermore, it effectively discovers critical microregions related to physiological and pathological variations offering new insight into metabolic heterogeneity.

Data reduction strategy in the PandaX-4T experiment

The PandaX-4T experiment is designed for multiple purposes, including searches for solar neutrinos, weakly interacting massive particles, and rare double beta decays of xenon isotopes. The experiment produces a huge amount of raw data that needs to be stored for related physical analyses in a wide energy range. With the upgrading of the PandaX-4T experiment, the doubled sampling rate resulted in a larger data size, which challenges both the cost and the data processing speed. To address this issue, we propose a data reduction strategy by removing the noise tail of large signals and downsampling the remaining parts of them. This strategy reduces the requirement for storage by 65% while increasing data processing speed. The influences on physical analyses on different topics at different energy regions are negligible.

Exploring the Performance of Linear and Nonlinear Models of Time-of-Flight Secondary Ion Mass Spectrometry Spectra.

Multivariate statistical tools and machine learning (ML) techniques can deconvolute hyperspectral data and control the disparity between the number of samples and features in materials science. Nevertheless, the importance of generating sufficient high-quality sample replicates in training data cannot be overlooked, as it fundamentally affects the performance of ML models. Here, we present a quantitative analysis of time-of-flight secondary ion mass spectrometry (ToF-SIMS) spectra of a simple microarray system of two food dyes using partial least-squares (PLS, linear) and random forest (RF, nonlinear) algorithms. This microarray was generated by a high-throughput sample preparation and analysis workflow for fast and efficient acquisition of quality and reproducible spectra via ToF-SIMS. We drew insights from the bias-variance trade-off, investigated the performances of PLS and RF regression models as a function of training data size, and inferred the amount of data needed to construct accurate and reliable regression models. In addition, we found that the spectral concatenation of positive and negative ToF-SIMS spectra improved the model performances. This study provides an empirical basis for future design of high-throughput microarrays and multicomponent systems, for the purpose of analysis with ToF-SIMS and ML.

Improving Security on Election Results Data Transmission via Cloud Using Hybrid Homomorphic Encryption

Elections in recent years have become topical issues and characterized by violence and conflicts leading to loss of lives and properties. The integrity and confidentiality of the declared collated results have been questioned by individuals and organisations with keen interest in the outcomes of every election. Different countries have adopted different Election Results Management Systems (RMS) to help present a credible, fair and transparent election results. These systems adopted are not without criticisms and suspicions. This research paper has presented various factors that need to be considered when selecting a Results Management System (RMS) for elections. A cloud-based using a Hybrid homomorphic encryption approach is proposed in managing the election results data security and transmission. The proposed scheme has demonstrated effectiveness in handling data integrity, data confidentiality, data privacy and access control. The proposed scheme presented has enhanced the security of election results data against Chosen Ciphetext Attacks (CCA) and Denial of Service (DoS) as well as other cyber related attacks. The time required for the entire election data encryption, transmission, decryption, upload time and download time has been greatly enhanced with the proposed system. The proposed system workflow algorithm, key generation algorithm, encryption algorithm and decryption have been presented in this research paper. The outcome of the research work indicates that only 0.00078 seconds is required to generate keys for about 100 users. About 0.705 seconds and 0.863 seconds is required for the encryption and decryption of a 500MB election results data. It was observed that the overall election results data transmission time is about 51.779 seconds which is less than one minute (60 seconds) for about 500MB data size of the election results data. This paper makes a case for the adoption and implementation of the proposed system since it performs better in terms of securing the election results data and transmission time in the cloud environment.