Sampling Scenarios Research Articles

Model predictive control of nonlinear processes using neural ordinary differential equation models

Neural Ordinary Differential Equation (NODE) is a recently proposed family of deep learning models that can perform a continuous approximation of a linear/nonlinear dynamic system using time-series data by integrating the neural network model with classical ordinary differential equation solvers. Modeling nonlinear dynamic processes using data has historically been a critical challenge in the field of chemical engineering research. With the development of computer science and neural network technology, recurrent neural networks (RNN) have become a popular black-box approach to accomplish this task and have been utilized to design model predictive control (MPC) systems. However, as a discrete-time approximation model, RNN requires strictly uniform step sequence data to operate, which makes it less robust to an irregular sampling scenario, such as missing data points during operation due to sensor failure or other types of random errors. This paper aims to develop a NODE model and construct an MPC based on this novel continuous-time neural network model. In this context, closed-loop stability is established and robustness to noise is addressed using a variety of data analysis techniques. An example of a chemical process is utilized to evaluate the performance of NODE-based MPC. Furthermore, the performance of the NODE-based MPC under Gaussian and non-Gaussian noise is investigated, and the subsampling method is found to be effective in building models for MPC that are suitable for handling the presence of non-Gaussian noise in the data.

A Comparative Study of Contrastive Learning-Based Few-Shot Unsupervised Algorithms for Efficient Deep Learning

The rapid development in computer vision algorithms, particularly in target detection algorithms based on deep neural networks, has led to significant advancements in various fields. However, recent research on target detection algorithms has shifted towards small sample scenarios with long-tailed distribution of categories, where achieving high accuracy target detection with limited data has become a crucial research topic. Deep neural networks often face several challenges when dealing with limited data, including convergence problems, overfitting, and poor generalization performance. In such scenarios, categories with little data are easily overwhelmed by the negative gradients generated by other categories during network training, which affects the final detection results. To address this issue, an unsupervised contrast learning algorithm has been proposed that can achieve accurate results on a small number of reference datasets without requiring a large number of datasets. One recently proposed contrast learning algorithm is PixelCL, which combines information from image depth in pretraining to improve results. The results of this algorithm show that it can obtain more accurate results through contrast learning with limited sample data. These findings highlight the potential of contrast learning algorithms in addressing negative gradient issues and improving the performance of deep neural networks in target detection tasks with limited data.

A Universal Feature Extractor Based on Self-Supervised Pre-Training for Fault Diagnosis of Rotating Machinery under Limited Data

To address the limited data problem in real-world fault diagnosis, previous studies have primarily focused on semi-supervised learning and transfer learning methods. However, these approaches often struggle to obtain the necessary data, failing to fully leverage the potential of easily obtainable unlabeled data from other devices. In light of this, this paper proposes a novel network architecture, named Signal Bootstrap Your Own Latent (SBYOL), which utilizes unlabeled vibration signals to address the challenging issues of variable working conditions, strong noise, and limited data in rotating machinery fault diagnosis. The architecture consists of a self-supervised pre-training-based fault feature recognition network and a diagnosis network based on knowledge transfer. The fault feature recognition network uses ResNet-18 as the backbone network for self-supervised pre-training and transfers the trained fault feature extractor to the target diagnostic object. Additionally, a unique vibration signal data augmentation technique, time–frequency signal transformation (TFST), is proposed specifically for rotating machinery fault diagnosis, which addresses the key task of contrastive learning and achieves high-precision fault diagnosis with very few labeled samples. Experimental results demonstrate that the proposed diagnostic model outperforms other methods in both extremely limited sample and strong noise scenarios and can transfer unlabeled data utilization between similar and even different device types.

Health estimation of lithium-ion batteries with voltage reconstruction and fusion model

Accurate state of health (SOH) estimation is crucial for Lithium-ion battery in electric vehicles (EVs). This work proposes a battery SOH estimation method based on voltage reconstruction and fusion models. Firstly, a voltage curve reconstruction method based on importance sampling is proposed to solve the V–Q curve. Then, feature factors related to SOH are extracted and their correlation with SOH is analyzed. Furthermore, a SOH estimation fusion model is established based on improved Support Vector Regression (SVR) and Convolutional Neural Network (CNN). Finally, the accuracy of the proposed algorithm is verified in 20% and 30% small sample scenarios, respectively. The experimental results show that the numerical evaluation indicators of the proposed method are superior to Gauss Process Regression (GPR), CNN, whale optimization algorithm-SVR (WOA-SVR) and Long Short Term Memory (LSTM) neural network, which indicates that the proposed method has good performance.

Least squares solvers for ill-posed PDEs that are conditionally stable

This paper is concerned with the design and analysis of least squares solvers for ill-posed PDEs that are conditionally stable. The norms and the regularization term used in the least squares functional are determined by the ingredients of the conditional stability assumption. We are then able to establish a general error bound that, in view of the conditional stability assumption, is qualitatively the best possible, without assuming consistent data. The price for these advantages is to handle dual norms which reduces to verifying suitable inf-sup stability. This, in turn, is done by constructing appropriate Fortin projectors for all sample scenarios. The theoretical findings are illustrated by numerical experiments.

A novel gridless space-time adaptive processing method based cyclic minimization

Space-time adaptive processing (STAP) algorithm based on sparse recovery (SR) can effectively suppress clutter in limited sample scenarios. However, the clutter suppression performance will be reduced because of grid mismatch (off-grid), which happens when the clutter ridge fails to match the dictionary. To address this issue, a novel gridless STAP algorithm based on cyclic minimization is proposed in this paper. Compared with the STAP algorithm based on atomic norm minimization (ANM), the proposed algorithm avoids the approximate loss caused by relaxing the rank norm to the nuclear norm. In addition, in order to improve the computational efficiency, we derive a fast iterative structure based on the alternating direction method of multipliers (ADMM) framework. The proposed algorithm offers better clutter suppression performance and excellent computational efficiency, according to a vast collection of experiments based on simulation data and measured data.

Predicting in-hospital mortality after transcatheter aortic valve replacement using administrative data and machine learning

Transcatheter aortic valve replacement (TAVR) is the gold standard treatment for patients with symptomatic aortic stenosis. The utility of existing risk prediction tools for in-hospital mortality post-TAVR is limited due to two major factors: (a) the predictive accuracy of these tools is insufficient when only preoperative variables are incorporated, and (b) their efficacy is also compromised when solely postoperative variables are employed, subsequently constraining their application in preoperative decision support. This study examined whether statistical/machine learning models trained with solely preoperative information encoded in the administrative National Inpatient Sample database could accurately predict in-hospital outcomes (death/survival) post-TAVR. Fifteen popular binary classification methods were used to model in-hospital survival/death. These methods were evaluated using multiple classification metrics, including the area under the receiver operating characteristic curve (AUC). By analyzing 54,739 TAVRs, the top five classification models had an AUC ≥ 0.80 for two sampling scenarios: random, consistent with previous studies, and time-based, which assessed whether the models could be deployed without frequent retraining. Given the minimal practical differences in the predictive accuracies of the top five models, the L2 regularized logistic regression model is recommended as the best overall model since it is computationally efficient and easy to interpret.

Development of dynamic integrated consequence evaluation (DICE) for dynamic event tree approaches: Numerical validation for a loss of coolant accident

Although conventional probabilistic safety assessment (PSA) makes a sufficient contribution to evaluating the safety of nuclear facilities, it suffers some limitations in performing optimal assessments close to the actual behavior by reflecting time dependencies in accident management. However, integrated deterministic and probabilistic safety assessment (IDPSA) literally combines two methodologies to enable synchronized evaluation based on physical behavior, system availability of facilities, and operator's response. Dynamic integrated consequence evaluation (DICE) is one of the IDPSA tools developed for this purpose, which is the first achievement in South Korea and supported by a regulatory authority. Based on dynamic discrete event tree (DDET) methodology, DICE consists of a scheduler that supports the exchange of information between modules, including a physical module that computes thermal-hydraulic simulations, a diagnosis module that specifies branching points for safety systems, and a reliability module that provides system availability. This paper deals with numerical validation comparing the conventional deterministic and probabilistic results for a sample scenario with those of DICE, and discusses an application of DICE to support and extend the scope of conventional PSA through a case study. It is expected that DICE emphasizes the compliance with conventional PSA methodologies while providing magnified perception in investigating unforeseen scenarios.

Light‐weight deep learning method for active jamming recognition based on improved MobileViT

AbstractIn recent years, there has been significant research focused on identifying jamming signals using neural networks to improve identification accuracy. However, traditional networks require large‐scale datasets for training, which is challenging in small sample scenarios such as actual battlefield jamming signal data. Additionally, the large number of parameters in traditional models presents deployment challenges for devices. To address these issues, this study proposes a lightweight neural network model called MobileViT_CA, an improved version of MobileViT, for jamming recognition in small sample scenarios. MobileViT_CA adopts a coordinate attention mechanism suitable for lightweight networks, which further improves the classification performance of the network. The model converts one‐dimensional jamming signals into two‐dimensional time‐frequency distribution images through time‐frequency analysis and then uses the network for recognition. Compared with traditional neural network models, MobileViT_CA significantly reduces the network parameter volume while maintaining high recognition accuracy, accelerating network inference speed, and reducing hardware requirements. Experimental results show that MobileViT_CA performs well in small sample data. The model's fast learning ability has also been verified by using transfer learning, and the use of Grad‐CAM and t‐SNE to analyse MobileViT_CA demonstrates its advantages in feature recognition and extraction. Overall, this approach has the potential to significantly improve the reliability and effectiveness of jamming recognition systems in battlefield scenarios.

Adversarial Autoencoder Empowered Joint Anomaly Detection and Signal Reconstruction From Sub-Nyquist Samples

Anomaly detection is an essential part of spectrum monitoring applications. Malicious users and malfunctioning nodes could be identified via anomaly detection methods. Meanwhile, the spectrum bands that would be utilized in future 6G or satellite communication system settings are going to be wider than ever. Acquiring Nyquist sampled data from such a spectrum would require components with a very high sampling rate. To monitor a wide spectrum, a compressive sensing recovery algorithm combined with a sub-sampling approach could accomplish the task with a lower hardware cost. To solve the anomaly detection problem using a sub-sampled data stream, a joint signal recovery and anomaly detection solution utilizing an adversarial autoencoder (AAE) structure are proposed in this article. An AAE is constructed via an autoencoder and a discriminator weaved together. The discriminator would guide the autoencoder to drive its extracted feature onto a designed feature space, while the autoencoder would provide a reconstruction of the original Nyquist sampled signal. The proposed AAE structure could learn the distribution of the signal from either labelled or unlabelled training data, enabling it to work on both supervised and unsupervised data sets. The proposed AAE has shown superior reconstruction and detection performance on very sparse sampling scenarios.

Simulation and experimental study of sediment sampling to improve sampling technique efficiency

Abyssal sediments are repositories of geological history and an important material basis for the study of marine biology, marine geology, paleoclimatology, and life evolution. Due to the interaction between sampler and sediments, the volumes of obtained sediment samples are commonly insufficient. Therefore, a sampling model and a numerical model calculation method are proposed to analyze the tubular sampling process in sediments. The sediments can be represented by non-Newtonian viscoplastic fluids, and sampling can be simulated using a dynamic grid technology in fluid calculation software. A three-dimensional flow field is dynamically simulated using the realizable k-ε turbulence model. The effects of structural parameters and working conditions on sample volume and sample rate were obtained. The results showed that when the drainage flow ratio is not greater than 3, drainage is sufficient. In order to improve sampling efficiency, the sample tube velocity should be maintained at 20 to 40 mm/s. By comparing the simulation to the laboratory sampling test and sea trials, absolute errors between 2.94% and 4.61% were revealed. This further verified that the sampling model can accurately reflect actual sampling scenarios. This findings provide a theoretical basis to optimize the structural design and parameters of samplers to achieve efficient sampling.

Timesweeper: accurately identifying selective sweeps using population genomic time series.

Despite decades of research, identifying selective sweeps, the genomic footprints of positive selection, remains a core problem in population genetics. Of the myriad methods that have been developed to tackle this task, few are designed to leverage the potential of genomic time-series data. This is because in most population genetic studies of natural populations, only a single period of time can be sampled. Recent advancements in sequencing technology, including improvements in extracting and sequencing ancient DNA, have made repeated samplings of a population possible, allowing for more direct analysis of recent evolutionary dynamics. Serial sampling of organisms with shorter generation times has also become more feasible due to improvements in the cost and throughput of sequencing. With these advances in mind, here we present Timesweeper, a fast and accurate convolutional neural network-based tool for identifying selective sweeps in data consisting of multiple genomic samplings of a population over time. Timesweeper analyzes population genomic time-series data by first simulating training data under a demographic model appropriate for the data of interest, training a one-dimensional convolutional neural network on said simulations, and inferring which polymorphisms in this serialized data set were the direct target of a completed or ongoing selective sweep. We show that Timesweeper is accurate under multiple simulated demographic and sampling scenarios, identifies selected variants with high resolution, and estimates selection coefficients more accurately than existing methods. In sum, we show that more accurate inferences about natural selection are possible when genomic time-series data are available; such data will continue to proliferate in coming years due to both the sequencing of ancient samples and repeated samplings of extant populations with faster generation times, as well as experimentally evolved populations where time-series data are often generated. Methodological advances such as Timesweeper thus have the potential to help resolve the controversy over the role of positive selection in the genome. We provide Timesweeper as a Python package for use by the community.

Security Ontology OntoSecRPA for Robotic Process Automation Domain

Robotic process automation (RPA)* based on the use of software robots has proven to be one of the most demanded technologies to emerge in recent years used for automating daily IT routines in many sectors, such as banking and finance. As with any new technology, RPA has a number of potential cyber security weaknesses, caused either by fundamental logical mistakes in the approach or by cyber-human mistakes made during the implementation, configuration, and operation phases. It is important to have an extensive understanding of the related risks before RPA integration into enterprise IT infrastructure. The main asset operated by RPA is confidential enterprise data. Data leakage and theft are the two main threats. The wide application of RPA technology in information security-sensitive sectors makes the protection of RPA against cyber-attacks an important task. Still, this topic is not yet adequately investigated in the scientific press and existing articles mainly concentrate on stating the RPA security importance and describing some threats. In this article, we present a flexible tool, security-oriented ontology OntoSecRPA*, which systematically describes RPA-specific assets, risks, security, threats, vulnerabilities, and countermeasures. To the best of our knowledge, there are currently no ontologies available that are specific to the RPA domain, and existing security ontologies lack RPA-related features. In the future, the proposed ontology can be updated and used in different ways, for example, as a checklist for risk management tasks in RPA solutions and a source of information for an expert system or a concentrated domain-specific source of information, which indicates its wide practical application. The proposed ontology was formally verified by applying ontology completeness assessment and used for risk assessment in a sample scenario.

Wind Retrieval from Constellations of Small SAR Satellites: Potential for Offshore Wind Resource Assessment

The planning of offshore wind energy projects requires wind observations over long periods for the establishment of wind speed distributions. In the marine environment, high-quality in situ observations are sparse and restricted to point locations. Numerical modeling is typically used to determine the spatial variability of the wind resource. Synthetic Aperture Radar (SAR) observations from satellites can be used for retrieval of wind fields over the ocean at a high spatial resolution. The recent launch of constellations of small SAR satellites by private companies will improve the sampling of SAR scenes significantly over the coming years compared with the current sampling rates offered by multi-purpose SAR missions operated by public space agencies. For the first time, wind fields are retrieved from a series of StriX SAR scenes delivered by Synspective (Japan) and also from Sentinel-1 scenes delivered by the European Space Agency. The satellite winds are compared with wind speed observations from the FINO3 mast in the North Sea. This leads to root-mean-square errors of 1.4–1.8 m s−1 and negative biases of −0.4 m s−1 and −1.0 m s−1, respectively. Although the Geophysical Model Functions (GMF) applied for wind retrievals have not yet been tuned for StriX SAR observations, the wind speed accuracy is satisfactory. Through conditional sampling, we estimate the wind resource from current and future SAR sampling scenarios where the number of SAR satellites in orbit is increasing over time. We find that hourly samples are needed to fully capture the diurnal wind speed variability at the site investigated. A combination of SAR samples from current missions with samples from clusters of small SAR satellites can yield the necessary number of wind speed samples for accurate wind resource estimation. This is particularly important for sites with pronounced diurnal wind speed variability. An additional benefit of small SAR satellites is that wind speed variability can be mapped at the sub-km scale. The very high spatial resolution is valuable for characterizing the wind conditions in the vicinity of existing offshore wind farms.

Developing solutions roadmaps for rice production in flooding area of the Vietnamese Mekong Delta adapting to uncertain changes in climate and surface water resources

Climate change, hydro-power dams and socio-economic development in the upper Mekong River, have caused changes in surface water resources and agricultural transformation in the Vietnamese Mekong Delta (VMD). It is unpredictable and influences decision-making processes on water resource management. This study aims at applying the DAPP (Dynamic Adaptive Policy Pathways) approach to develop solution roadmaps supporting rice production in the Mekong Delta projection to 2030 under the context of uncertain changes on surface water resources. Data were collected from local officers, farmers, experts at Can Tho University, Technical University Delft, and Utrecht University using semi-structured interview techniques including workshops, group discussions, and in-depth interviews. The consolidated solutions from stakeholder’s opinions were evaluated at different socio-economic and environmental dimensions, specially focused on the time of tipping point of each solution. The study results show that two solution roadmaps were developed as sample scenarios to promote rice production in the upper area of the VMD. The determination of the solutions roadmaps based on the agreement of the stakeholders, which is a new feature of this approach in linking many stakeholders for the general development of the region. The DAPP approach could be suitable to support decision-making on surface water resources management in the context of uncertain changes. Therefore, further studies can apply this approach to other regions in the VMD which will support to reinforce and demonstrate the applicability of the DAPP approach on supporting decision-making in the face of increasingly uncertain changes in the VMD.

Continuous Flow Separation of Red Blood Cells and Platelets in a Y-Microfluidic Channel Device with Saw-Tooth Profile Electrodes via Low Voltage Dielectrophoresis

Cell counting and sorting is a vital step in the purification process within the area of biomedical research. It has been widely reported and accepted that the use of hydrodynamic focusing in conjunction with the application of a dielectrophoretic (DEP) force allows efficient separation of biological entities such as platelets from red blood cell (RBC) samples due to their size difference. This paper presents computational results of a multiphysics simulation modelling study on evaluating continuous separation of RBCs and platelets in a microfluidic device design with saw-tooth profile electrodes via DEP. The theoretical cell particle trajectory, particle cell counting, and particle separation distance study results reported in this work were predicted using COMSOL v6.0 Multiphysics simulation software. To validate the numerical model used in this work for the reported device design, we first developed a simple y-channel microfluidic device with square "in fluid" electrodes similar to the design reported previously in other works. We then compared the obtained simulation results for the simple y-channel device with the square in fluid electrodes to the reported experimental work done on this simple design which resulted in 98% agreement. The design reported in this work is an improvement over existing designs in that it can perform rapid separation of RBCs (estimated 99% purification) and platelets in a total time of 6-7 s at a minimum voltage setting of 1 V and at a minimum frequency of 1 Hz. The threshold for efficient separation of cells ends at 1000 kHz for a 1 V setting. The saw-tooth electrode profile appears to be an improvement over existing designs in that the sharp corners reduced the required horizontal distance needed for separation to occur and contributed to a non-uniform DEP electric field. The results of this simulation study further suggest that this DEP separation technique may potentially be applied to improve the efficiency of separation processes of biological sample scenarios and simultaneously increase the accuracy of diagnostic processes via cell counting and sorting.

Designing a Long-term Occupancy Monitoring Plan for a Cryptic Reptile

Monitoring populations of cryptic reptiles is challenging because of their crypticity. Occupancy monitoring is a useful technique for local populations, but seasonal unavailability for detection can result in large swings in apparent annual occupancy. We used data from 5 yr of occupancy surveys, and the observed sampling and process error, to evaluate the power to detect true change in occupancy under a range of sampling scenarios and occupancy trends for a cryptic reptile, Mojave Desert Tortoise (Gopherus agassizii) None of the sampling strategies that we tested had the power to detect a 1% per annum increase in true occupancy over a 10-yr period and required ≥70 sites to detect a 1% decline in occupancy. For a 2% per annum change in true occupancy, 50 sites were needed and for a 3–4% per annum change 20–30 sites were needed. Power to detect a 4% per annum decline in occupancy was >80% for all number of secondary visits and number of sites if the number of sites was ≥50, but required seven visits to ≥60 sites to detect a 4% increase in occupancy Multiple sampling designs provided equivalent power. For example, both designs of 80 sites with three secondary visits and 50 sites with seven secondary visits yielded >50% power on average to detect 1–4% per annum changes in true occupancy. These results can help local managers of Mojave Desert Tortoises and other cryptic reptiles to design optimally efficient occupancy monitoring strategies.

Vibration-based incipient surge detection and diagnosis of the centrifugal compressor using adaptive feature fusion and sparse ensemble learning approach

As a critical high-speed rotating machinery, the centrifugal compressor has been widely used in various modern industries. However, it is subject to a potential damaging phenomenon called surge that can be caused by several factors such as unmatched design, improper operation, inlet and outlet blockage, and so on, which thus may result in catastrophic accidents. This paper concerns the incipient surge detection and diagnosis (ISDD) of the centrifugal compressor based on its bearing vibration signals, leveraging adaptive feature fusion and sparse ensemble learning approach to develop a data-driven-based intelligent diagnostic model. Firstly, vibration signals are decomposed by the empirical mode decomposition (EMD) method, and various features can be extracted from the obtained intrinsic mode functions (IMFs) with rich information of the time series. Next, maximum likelihood estimation (MLE) is utilized to compute the intrinsic dimension of the extracted feature vectors, which are then reduced to the corresponding dimensionality adaptively by kernel principal component analysis (kernel PCA). Coping with the multi-dimensional input and nonlinearly separable classification problem, an ensemble learning approach with L1-regularization term is adopted for incipient surge identification of the centrifugal compressor. A novel and hyperparameter-relaxed hybrid algorithm is employed for the sparse ensemble learning procedure, which guarantees transcendental accuracy and preeminent generalization capability while ensuring high robustness and sparsity of the intelligent model. Finally, in addition to diagnostic tasks of the centrifugal compressor datasets under different working conditions, the ISDD performance of the proposed method in imbalanced sample scenarios, varying working conditions and additional noise interference circumstances is comprehensively investigated, and the overall success indicates the robustness and superiority of the proposed method compared with other existing approaches.

Homogeneity Test of the First-Order Agreement Coefficient in a Stratified Design

Gwet's first-order agreement coefficient (AC1) is widely used to assess the agreement between raters. This paper proposes several asymptotic statistics for a homogeneity test of stratified AC1 in large sample sizes. These statistics may have unsatisfactory performance, especially for small samples and a high value of AC1. Furthermore, we propose three exact methods for small pieces. A likelihood ratio statistic is recommended in large sample sizes based on the numerical results. The exact E approaches under likelihood ratio and score statistics are more robust in the case of small sample scenarios. Moreover, the exact E method is effective to a high value of AC1. We apply two real examples to illustrate the proposed methods.

Untangling the Sources of Abundance Dispersion in Low-metallicity Stars

We measure abundances of 12 elements (Na, Mg, Si, Ca, Sc, Ti, V, Cr, Mn, Fe, Co, Ni) in a sample of 86 metal-poor (−2 ≲ [Fe/H] ≲ −1) subgiant stars in the solar neighborhood. Abundances are derived from high-resolution spectra taken with the Potsdam Echelle Polarimetric and Spectroscopic Instrument on the Large Binocular Telescope, modeled using iSpec and MOOG. By carefully quantifying the impact of photon-noise (<0.05 dex for all elements), we robustly measure the intrinsic scatter of abundance ratios. At fixed [Fe/H], the rms intrinsic scatter in [X/Fe] ranges from 0.04 (Cr) to 0.16 dex (Na), with a median of 0.08 dex. Scatter in [X/Mg] is similar, and accounting for [α/Fe] only reduces the overall scatter moderately. We consider several possible origins of the intrinsic scatter with particular attention to fluctuations in the relative enrichment by core-collapse supernovae (CCSN) and Type Ia supernovae and stochastic sampling of the CCSN progenitor mass distribution. The stochastic sampling scenario provides a good quantitative explanation of our data if the effective number of CCSN contributing to the enrichment of a typical sample star is N ∼ 50. At the median metallicity of our sample, this interpretation implies that the CCSN ejecta are mixed over a gas mass ∼6 × 104 M ⊙ before forming stars. The scatter of elemental abundance ratios is a powerful diagnostic test for simulations of star formation, feedback, and gas mixing in the early phases of the Galaxy.