Multiple Vectors Research Articles

Prediction of the health state of lithium-ion power batteries using machine learning

Faced with the increasingly urgent issues of energy depletion and environmental pollution, various countries are actively supporting the new energy vehicle industry, making electric vehicle technology a focus of widespread attention among researchers. The Battery Management System (BMS) plays a crucial role in overseeing the power battery's operational parameters, with the estimation of the State of Health (SOH) being a pivotal function. Accurate estimation of SOH can improve the utilization efficiency and endurance of power batteries, extend their service life, and help users achieve the best balance between system safety and economic benefits. Machine learning methods provide a new solution to the SOH estimation problem. Without analyzing the complex aging mechanisms inside the battery, online SOH estimation can be completed by learning from historical aging data. In this study, the lithium battery dataset provided by NASA is used and trained through multiple linear regression models and support vector machine models to predict the SOH of lithium-ion batteries. The results demonstrate the accuracy and reliability of both methods in predicting SOH. Simultaneously, the prediction results of different feature values are compared to obtain the most accurate combination of feature values.

Simultaneous Expression of Different Therapeutic Genes by Infection with Multiple Oncolytic HSV-1 Vectors.

Oncolytic viruses (OVs) are anti-cancer therapeutics combining the selective killing of cancer cells with the triggering of an anti-tumoral immune response. The latter effect can be improved by arming OVs with immunomodulatory factors. Due to the heterogeneity of cancer and the tumor microenvironment, it is anticipated that strategies based on the co-expression of multiple therapeutic molecules that interfere with different features of the target malignancy will be more effective than mono-therapies. Here, we show that (i) the simultaneous expression of different proteins in triple-negative breast cancer (TNBC) cells can be achieved through their infection with a combination of OVs based on herpes simplex virus type 1 (oHSV1), each encoding a single transgene. (ii) The level of expressed proteins is dependent on the number of infectious viral particles utilized to challenge tumor cells. (iii) All recombinant viruses exhibited comparable efficacy in the killing of TNBC cells in single and multiple infections and showed similar kinetics of replication. Overall, our results suggest that a strategy based on co-infection with a panel of oHSV1s may represent a promising combinatorial therapeutic approach for TNBC, as well as for other types of solid tumors, that merits further investigation in more advanced in vitro and in vivo models.

Bluetongue Risk Map for Vaccination and Surveillance Strategies in India.

Bluetongue virus (BTV, Sedoreoviridae: Orbivirus) causes an economically important disease, namely, bluetongue (BT), in domestic and wild ruminants worldwide. BTV is endemic to South India and has occurred with varying severity every year since the virus was first reported in 1963. BT can cause high morbidity and mortality to sheep flocks in this region, resulting in serious economic losses to subsistence farmers, with impacts on food security. The epidemiology of BTV in South India is complex, characterized by an unusually wide diversity of susceptible ruminant hosts, multiple vector species biting midges (Culicoides spp., Diptera: Ceratopogonidae), which have been implicated in the transmission of BTV and numerous co-circulating virus serotypes and strains. BT presence data (1997-2011) for South India were obtained from multiple sources to develop a presence/absence model for the disease. A non-linear discriminant analysis (NLDA) was carried out using temporal Fourier transformed variables that were remotely sensed as potential predictors of BT distribution. Predictive performance was then characterized using a range of different accuracy statistics (sensitivity, specificity, and Kappa). The top ten variables selected to explain BT distribution were primarily thermal metrics (land surface temperature, i.e., LST, and middle infrared, i.e., MIR) and a measure of plant photosynthetic activity (the Normalized Difference Vegetation Index, i.e., NDVI). A model that used pseudo-absence points, with three presence and absence clusters each, outperformed the model that used only the recorded absence points and showed high correspondence with past BTV outbreaks. The resulting risk maps may be suitable for informing disease managers concerned with vaccination, prevention, and control of BT in high-risk areas and for planning future state-wide vector and virus surveillance activities.

Parallel vector memories in the brain of a bee as foundation for flexible navigation

Insects rely on path integration (vector-based navigation) and landmark guidance to perform sophisticated navigational feats, rivaling those seen in mammals. Bees in particular exhibit complex navigation behaviors including creating optimal routes and novel shortcuts between locations, an ability historically indicative of the presence of a cognitive map. A mammalian cognitive map has been widely accepted. However, in insects, the existence of a centralized cognitive map is highly contentious. Using a controlled laboratory assay that condenses foraging behaviors to short distances in walking bumblebees, we reveal that vectors learned during path integration can be transferred to long-term memory, that multiple such vectors can be stored in parallel, and that these vectors can be recalled at a familiar location and used for homeward navigation. These findings demonstrate that bees meet the two fundamental requirements of a vector-based analog of a decentralized cognitive map: Home vectors need to be stored in long-term memory and need to be recalled from remembered locations. Thus, our data demonstrate that bees possess the foundational elements for a vector-based map. By utilizing this relatively simple strategy for spatial organization, insects may achieve high-level navigation behaviors seen in vertebrates with the limited number of neurons in their brains, circumventing the computational requirements associated with the cognitive maps of mammals.

Parallel ensemble of a randomization-based online sequential neural network for classification problems using a frequency criterion

Randomization-based neural networks have gained wide acceptance in the scientific community owing to the simplicity of their algorithm and generalization capabilities. Random vector functional link (RVFL) networks and their variants are a class of randomization-based neural networks. RVFL networks have shown promising results in classification, regression, and clustering problems. For real-world applications, learning algorithms that can train with new samples over previous results are necessary because of to the constant generation of problems related to large-scale datasets. Various online sequential algorithms, commonly involving an initial learning phase followed by a sequential learning phase, have been proposed to address this issue. This paper presents a training algorithm based on multiple online sequential random vector functional link (OS-RVFL) networks for large-scale databases using a shared memory architecture. The training dataset is distributed among p OS-RVFL networks, which are trained in parallel using p threads. Subsequently, the test dataset samples are classified using each trained OS-RVFL network. Finally, a frequency criterion is applied to the results obtained from each OS-RVFL network to determine the final classification. Additionally, an equation was derived to reasonably predict the total training time of the proposed algorithm based on the learning time in the initial phase and the time scaling factor compared to the sequential learning phase. The results demonstrate a drastic reduction in training time because of data distribution and an improvement in accuracy because of the adoption of the frequency criterion.

Single-Frame Lagrangian Tracking Of 3-D Acoustic Streaming Flows Using Digital Defocusing Micro Particle Streak Velocimetry

In this study, a single-frame particle streak velocimetry technique is applied to the classic DDPIV(Digital Defocusing Particle Image Velocimetry) concept to achieve a single-frame 3-D Lagrangian tracking of particles in microscopic flows. The shortcoming of images suffering low signal to noise ratio due to pinholes can be avoided. With the streak resolving algorithm, it allows extended exposure time when streak images are taken, and the single-frame approach also eliminates the need for high-cost and complicated synchronization hardware for double-pulse frame-straddling light sources. For each long-exposure, color-coded images received from each pinhole equipped with red, green or blue color filters, the streak-resolving algorithm is applied first to find each streak on the separate color channel, followed by an one-time triplet-matching process to group the red, green and blue streak projections. From the streak triplets, 3-D reconstruction can be done on all the resolved points on the streak, and trajectory fitting can then be applied to resolve the tracer particle trajectories in the field of view with temporal history. This procedure was applied for visualizing a microscale 3-D acoustic streaming flow pattern induced by a longitudinal spine-shaped fin oscillating at 12kHz. The height of the channel is 1mm with the spine of 0.3mm height and a 30 。 tip angle. Resolved flow fields show that the particle streak images can be resolved with the same level of accuracy to the particle tracking method, while the throughput of velocity vectors can be significantly higher due to multiple velocity vectors can be resolved from each streak triplet. The methodology has the potential to be applied to various applications that can capture long-exposure images and possible to resolve higher-order information such as acceleration and forces applied on the particles or the flow.

Multiplicative rectifying submanifolds of multiplicative Euclidean space

In this paper, we initiate the study of the multiplicative Euclidean submanifolds, exhibiting the multiplicative counterparts of the Gauss–Weingarten formulas. Some particular types such as multiplicative conic, spherical, and totally geodesic submanifolds are analyzed. In relation to such multiplicative submanifolds, we introduce multiplicative rectifying submanifolds and characterize by means of multiplicative concurrent vector fields. The multiplicative rectifying hypersurfaces are also obtained.

Multi-Interest Sequential Recommendation with Simplified Graph Convolution and Multiple Item Features

Multi-interest sequential recommendations leverage users’ historical behavior to provide recommendations that match multiple interests. Most of these methods have not fully extracted higher-order information hidden in users’ interactions and have overlooked the multiple features of items. To this end, this paper proposes a multi-interest model called “multi-interest sequential recommendation with simplified graph convolution and item multi-features (SGCMF)”. Firstly, a simplified graph convolution module is designed based on bipartite graphs, which utilizes mean pooling to aggregate neighboring information and employs a feedforward neural network (FNN) for nonlinear transformations and combinations. This method reduces redundant information and captures higher-order relationships, thereby simplifying the complexity of modeling high-order interactions and improving prediction accuracy. Secondly, an item multi-feature extraction module is proposed, which represents item features with multiple vectors, and analyzes each feature from multiple perspectives while preserving important relationships between features. The model correlates multiple features of the item with user interests, thereby achieving a fine-grained analysis of user interests. Extensive experiments are conducted on five real-world scenarios, and the results are compared with state-of-the-art methods. The experimental results show that SGCMF outperforms other baselines.

Lagrange Multipliers Aided Modulated Model Predictive Control Technique for PMSM Drives

The modulated model predictive control (MMPC) method utilizes multiple voltage vectors in sequence within one switching cycle, achieving a fixed switching frequency. However, the control performance of using the MMPC technique still suffers from imprecise vector duration derivation and high computational burden. To address this, this paper proposes an improved MMPC technique aided by the Lagrange multiplier method to accurately calculate the duration of each vector, and then efficiently select the optimal voltage vector. Compared to the existing MMPC techniques, the proposed one exhibits a lower computational burden and superior steady‐state performance. Those benefits have been proven by the experimental results of a 1.5 k rpm permanent magnet synchronous motor drive. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Machine learning-based aggressiveness assessment model construction for crabs: A case study of swimming crab Portunus trituberculatus

Aggressiveness trait-based selection is crucial for alleviating interspecies cannibalism in economic crab species and enhancing survival rates in aquaculture. However, there is a lack of efficient and simple methods for assessing aggressiveness. In this study, we measured aggressiveness of the swimming crab Portunus trituberculatus through repeated mirror tests and fighting experiments. Factor analysis and the K-means algorithm were used to assess aggressiveness quantitatively and qualitatively. A combination of multiple linear regression and support vector machine (SVM) analyses was employed to construct an aggressiveness assessment model for swimming crabs and explore the relationship between aggressiveness and fighting ability. The results showed significant correlations among repeated aggressive behaviors (attacking, chela extending, defending, crossing, reverse walking, and freezing). Aggression score was significantly correlated with fighting behaviors, and there were significant differences in fighting abilities among different levels of aggressiveness. This suggested that aggressive behaviors are consistent within individuals and that aggressiveness, as a personal trait, affects the fighting ability of swimming crabs. Aggression score (Y) and clustering results of K-means can serve as assessment indicators of aggressiveness. The predictive variables for the quantitative assessment model were relative movement distance (X1) and freezing duration (X2). The adjusted R-square of the optimized quantitative model was 0.72, it also had the smallest Sigma, AIC, MSE, and RMSE values and the best fitting regression equation, which was Y = 0.023X1 – 0.001X2 – 0.002. The predictor variables for the qualitative assessment model were relative movement distance, freezing frequency, and duration. SVM was used to construct the qualitative model, and the prediction accuracy was 92%, sensitivity was 84%, and specificity was 100%, indicating the model has a good classification and prediction effect. The machine learning-based aggressiveness assessment model constructed in this study provides a behavioral method for the selection and high-throughput measurement of economic crab species with excellent aggressiveness traits, giving it important industrial application value.

Sampling and active learning methods for network reliability estimation using K-terminal spanning tree

Network reliability analysis remains a challenge due to the increasing size and complexity of networks. This paper presents a novel sampling method and an active learning method for efficient and accurate network reliability estimation under node failure and edge failure scenarios. The proposed sampling method adopts Monte Carlo technique to sample component lifetimes and the K-terminal spanning tree algorithm to accelerate structure function computation. Unlike existing methods that compute only one structure function value per sample, our method generates multiple component state vectors and corresponding structure function values from each sample. Network reliability is estimated based on survival signatures derived from these values. A transformation technique extends this method to handle both node failure and edge failure. To enhance efficiency of proposed sampling method and achieve adaptability to network topology changes, we introduce an active learning method utilizing a random forest (RF) classifier. This classifier directly predicts structure function values, integrates network behaviors across diverse topologies, and undergoes iterative refinement to enhance predictive accuracy. Importantly, the trained RF classifier can directly predict reliability for variant networks, a capability beyond the sampling method alone. Through investigating several network examples and two practical applications, the effectiveness of both proposed methods is demonstrated.

Calibrating lab and field reflectance spectra for nutrient estimation in potato plants using local support vector regression models

This study presents a methodology based on multiple local support vector regression (SVR) to calibrate the spectra taken in the field in relative to lab-derived spectra. Laboratory based foliar spectral measurement is a common method to provide lab-derived spectra as a service where a grower sends sample leaves collected manually. The drawback of this method is being time-consuming when the samples are collected and analyzed. In contrast, in-field spectral measurements can be an alternative method capable of providing immediate readings. While both methods work based on the same priniciple, the insturmental differences as well as the conditional difference under which the instruments operate may cause differences in the spectral patterns of the same target. In this work, after developing the calibration method, we validated it by estimating NPK measurements in potato plants using in-field, lab, and field calibrated spectral measurements over two testing modes: dried and fresh. The results showed that the calibration using SVR models could minimize the percentage relative error (PRE) between lab and field spectra within the visible range by considering the influence of the neighboring wavebands up to 32 nm width which improved the alignment of the local maxima of the specral curves. Also, a substantial PRE reduction from 120 % to 20 % for some wavebands in the short-wave infrared (SWIR) region of the fresh mode was observed due to the influence of scaling within the SVR method. The calibration improved the alignment of NPK estimated values between lab and field calibrated spectra of both modes with an emphasis on its necessity to estimate nutrients in the fresh mode as the root mean square error was < 0.1 for the three elements.

Experimental investigation on the deformation behavior of an isotropic 304L austenitic steel manufactured by laser powder bed fusion with hot isostatic pressing

Post-heat treatment, such as hot isostatic pressing (HIP), is increasingly employed in laser powder bed fused (L-PBF) metallic materials to release residual stress, alleviate microstructural anisotropy, and tailor mechanical properties. Here, we systematically investigate the initial and deformed microstructures of as-received and HIPed L-PBF 304L austenitic steel using various microstructural characterizations. The results indicate that the HIP procedure can decrease the porosity and generate the columnar-to-equiaxed transition (CET), resulting in an isotropic microstructural morphology better suitable for practice applications. Additionally, the HIP samples have exceptional ductility when compared to as-received samples, which is associated with the sustained strain hardening rate arising from the notable deformation twinning behavior. The majority of deformation twins are preferentially emitted from the straight annealing twin boundaries and then grow within the 〈111〉 − oriented annealing twins. The principle twinning mechanism involved is the twinning partial emission from the annealing twin boundaries or high-angle grain boundaries (HAGBs). Moreover, the inclusions distributed near/on the annealing twin boundaries can emit partials with multiple Burgers vectors during straining, resulting in the extra random activation of partials (RAP) twinning mechanism that sustains the deformation process. These findings give helpful suggestions for HIP process designs of L-PBF austenitic steels, as well as a solid foundation for future theoretical modeling research.

Multi-step ahead forecasting of wind vector for multiple wind turbines based on new deep learning model

Obtaining wind speed and direction predictions with high accuracy is of vital significance for the utilization of wind energy. This research focuses on three problems: how to realize multiple station prediction, multi-step ahead prediction and wind vector prediction. Firstly, a 4-D time series variable scene is proposed for multi-step ahead forecasting of multiple station wind vector. Then, a wind vector prediction model (CM-G) based on Convolutional Minimum Gate Memory Neural Network and Graph Convolution Neural Network is constructed to improve the prediction accuracy. Further, three multi-step ahead forecasting modes are compared and their advantages, disadvantages and application scenarios are analyzed. Finally, the model and method proposed in this study are applied to two datasets in Tibet, China. The experimental results and conclusions are as follows: (1) The prediction accuracy of CM-G proposed in this study is higher than that of the existing state-of-the-art model in 90 % of the metrics. (2) From single-step to multi-step ahead forecasting, the prediction accuracy of wind vector gradually decreases. (3) In wind vector prediction, the prediction accuracy of wind speed is higher than that of wind direction. (4) The direction of accuracy improvement for multiple stations is relatively consistent with the average wind direction.

A variational Crank–Nicolson ensemble Monte Carlo algorithm for a heat equation under uncertainty

In this paper, a new fully discrete variational Crank–Nicolson ensemble (FVE/CN) scheme is established for the heat equation with uncertainty, then a fully discrete variational Crank–Nicolson ensemble Monte Carlo algorithm (FVEMC/CN) is given. Using the FVEMC/CN algorithm, one just needs to solve a single linear system with multiple right-hand side vectors in a group at each time step. The stability and the error estimates are presented to indicate the proposed algorithm is second-order convergent regarding the time. Finally, numerical experiments are provided which illustrate the convergence theory and the efficiency of the proposed approaches.

A Novel Approach to Image Retrieval for Vision-Based Positioning Utilizing Graph Topology

Abstract. This research introduces a novel approach to improve vision-based positioning in the absence of GNSS signals. Specifically, we address the challenge posed by obstacles that alter image information or features, making retrieving the query image from the database difficult. While the Bag of Visual Words (BoVW) is a widely used image retrieval technique, it has a limitation in representing each image with a single histogram vector or vocabulary of visual words, i.e., the emergence of obstacles can introduce new features to the query image, resulting in different visual words. Our study overcomes this limitation by clustering the features of each image using the k-means method and generating a graph for each class. Each node or key point in the graph obtains additional information from its direct neighbors using functions employed in graph neural networks, functioning as a feedforward network with constant parameters. This process generates new embedding nodes, and eventually, global pooling is applied to produce one vector for each graph, representing each image with graph vectors based on objects or feature classes. As a result, each image is represented with graph vectors based on objects or feature classes. In the presence of obstacles covering one or more graphs, there is sufficient information from the query image to retrieve the most relevant image from the database. Our approach was applied to indoor positioning applications, with the database collected in Bolz Hall at The Ohio State University. Traditional BoVW techniques struggle to properly retrieve most query images from the database due to obstacles like humans or recently deployed objects that alter image features. In contrast, our approach has shown progress in image retrieval by representing each image with multiple graph vectors, depending on the number of objects in the image. This helps prevent or mitigate changes in image features caused by obstacles covering or adding features to the image, as demonstrated in the results.

Modeling the dynamics of Onchocerca volvulus with the impact of environmental factors on blackfly breeding sites

Onchocerciasis is a significant public health concern, characterized by infections affecting the eyes and skin. The environmental factors contributing to its widespread occurrence are frequently undervalued, despite their significant influence. This research introduces an innovative mathematical model that utilizes fractional derivatives to examine the impact of several important variables, including temperature, river velocity, and oxygen concentration, on the breeding locations of blackflies. The objective of our study is to improve comprehension regarding the influence of environmental elements on the dynamics of onchocerciasis transmission. The efficacy of multiple vector control tactics and host screening approaches, such as skin snip biopsy and continuous detection at healthcare facilities, is evaluated through a series of comprehensive numerical experiments. Our results highlights the potential effectiveness of implementing comprehensive vector management strategies, in conjunction with frequent host screening techniques, in substantially decreasing the rates of disease transmission. Our study specifically emphasizes the efficacy of these interventions in riverine environments that are characterized by varying environmental circumstances. The findings obtained from this study have important significance for the development and execution of focused public health interventions aimed at combating onchocerciasis in areas where environmental factors are crucial in the transmission dynamics of the illness. This study provides significant insights into the worldwide endeavor to combat neglected tropical diseases, highlighting the critical importance of environmental factors and implementing integrated control techniques for efficient disease management and prevention.

Winter diet of bats in working forests of the southeastern U.S. Coastal Plain

Working forests comprise a large proportion of forested landscapes in the southeastern United States and are important to the conservation of bats, which rely on forests for roosting and foraging. While relationships between bat ecology and forest management are well studied during summer, winter bat ecology remains understudied. Hence, we aimed to identify the diet composition of overwintering bats, compare the composition of prey consumed by bat species, and determine the potential role of forest bats as pest controllers in working forest landscapes of the southeastern U.S. Coastal Plain. During January to March 2021–2022, we captured 264 bats of eight species. We used DNA metabarcoding to obtain diet composition from 126 individuals of seven bat species identifying 22 orders and 174 families of arthropod prey. Although Coleoptera, Diptera, and Lepidoptera were the most consumed orders, we found that bats had a generalist diet but with significant differences among some species. We also documented the consumption of multiple insect pests (e.g., Rhyacionia frustrana) and disease vectors (e.g., Culex spp). Our results provide important information regarding the winter diet of bats in the southeastern U.S. Coastal Plain and their potential role in controlling economically relevant pest species and disease vectors.

Evaluating Explainable Machine Learning Models for Clinicians

Gaining clinicians’ trust will unleash the full potential of artificial intelligence (AI) in medicine, and explaining AI decisions is seen as the way to build trustworthy systems. However, explainable artificial intelligence (XAI) methods in medicine often lack a proper evaluation. In this paper, we present our evaluation methodology for XAI methods using forward simulatability. We define the Forward Simulatability Score (FSS) and analyze its limitations in the context of clinical predictors. Then, we applied FSS to our XAI approach defined over an ML-RO, a machine learning clinical predictor based on random optimization over a multiple kernel support vector machine (SVM) algorithm. To Compare FSS values before and after the explanation phase, we test our evaluation methodology for XAI methods on three clinical datasets, namely breast cancer, VTE, and migraine. The ML-RO system is a good model on which to test our XAI evaluation strategy based on the FSS. Indeed, ML-RO outperforms two other base models—a decision tree (DT) and a plain SVM—in the three datasets and gives the possibility of defining different XAI models: TOPK, MIGF, and F4G. The FSS evaluation score suggests that the explanation method F4G for the ML-RO is the most effective in two datasets out of the three tested, and it shows the limits of the learned model for one dataset. Our study aims to introduce a standard practice for evaluating XAI methods in medicine. By establishing a rigorous evaluation framework, we seek to provide healthcare professionals with reliable tools for assessing the performance of XAI methods to enhance the adoption of AI systems in clinical practice.

Advancing language models through domain knowledge integration: a comprehensive approach to training, evaluation, and optimization of social scientific neural word embeddings

This article proposes a comprehensive strategy for training, evaluating, and optimizing domain-specific word2vec-based word embeddings, using social science literature as an example. Our primary objectives are: (1) to train the embeddings utilizing a corpus of social science text, (2) to test their performance against domain-unspecific embeddings using our developed intrinsic and extrinsic evaluation strategy, and (3) to enhance their performance even further by using domain knowledge. As an integral part of this approach, we present SociRel-461, a domain-knowledge dictionary designed for the intrinsic evaluation and subsequent refinement of social science word embeddings. Using a dataset of 100,000 full-text scientific articles in sociology, we train multiple vector space models, which we then benchmark against a larger, pre-trained general language embedding model as part of our extrinsic evaluation. Furthermore, we developed a transfer learning multi-label classification task for extrinsic evaluation. Our findings reveal that domain-specific embeddings outperform their domain-unspecific counterparts in both intrinsic and extrinsic evaluations. We also investigated the retrofitting post-processing method to enhance domain-unspecific embeddings with the domain knowledge embedded in SociRel-461. While retrofitting does not enhance our domain-specific vector space models, it significantly improves the performance of the domain-unspecific embeddings. This highlights the potential of retrofitting for the transfer of domain knowledge to domain-unspecific embeddings. Our results emphasize the importance of utilizing domain-specific word embeddings for better performance in domain specific transfer learning tasks, as they outperform conventional embeddings trained on everyday language.