A Parallel Multiclassification Algorithm for Big Data Using an Extreme Learning Machine

Controlled Migration of Lithium Cations by Diamine Bridges in Water-Processable Polymer-Based Solid-State Electrolyte Memory Layers for Organic Synaptic Transistors

Synaptic transistors require sufficient retention (memory) performances of current signals to exactly mimic biological synapses. Ion migration has been proposed to achieve high retention characteristics but less attention has been paid to polymer-based solid-state electrolytes (SSEs) for organic synaptic transistors (OSTRs). Here, OSTRs with water-processable polymer-based SSEs, featuring ion migration-controllable molecular bridges, which are prepared by reactions of poly(4-styrenesulfonic acid) (PSSA), diethylenetriamine (DETA), and lithium hydroxide (LiOH) are demonstrated. The ion conductivity of PSSA:LiOH:DETA (1:0.4:X, PLiD) films is remarkably changed by the molar ratio (X) of DETA, which is attributed to the extended distances between the PSSA chains by the DETA bridges. The devices with the PLiD layers deliver noticeably changed hysteresis reaching an optimum at X = 0.2, leading to the longest retention of current signals upon single/double pulses. The long-term potentiation test confirms that the present OSTRs can gradually build up the postsynaptic current by gate pulses of -2 V, while the long-term depression can be adjusted by varying the depression gate pulses (≈0.2-1.2 V). The artificial neural network simulations disclose that the present OSTRs with the ion migration-controlled PLiD layers can perform synaptic processes with an accuracy of ≈96%.

Skin Lesion Synthesis and Classification Using an Improved DCGAN Classifier

The prognosis for patients with skin cancer improves with regular screening and checkups. Unfortunately, many people with skin cancer do not receive a diagnosis until the disease has advanced beyond the point of effective therapy. Early detection is critical, and automated diagnostic technologies like dermoscopy, an imaging device that detects skin lesions early in the disease, are a driving factor. The lack of annotated data and class-imbalance datasets makes using automated diagnostic methods challenging for skin lesion classification. In recent years, deep learning models have performed well in medical diagnosis. Unfortunately, such models require a substantial amount of annotated data for training. Applying a data augmentation method based on generative adversarial networks (GANs) to classify skin lesions is a plausible solution by generating synthetic images to address the problem. This article proposes a skin lesion synthesis and classification model based on an Improved Deep Convolutional Generative Adversarial Network (DCGAN). The proposed system generates realistic images using several convolutional neural networks, making training easier. Scaling, normalization, sharpening, color transformation, and median filters enhance image details during training. The proposed model uses generator and discriminator networks, global average pooling with 2 × 2 fractional-stride, backpropagation with a constant learning rate of 0.01 instead of 0.0002, and the most effective hyperparameters for optimization to efficiently generate high-quality synthetic skin lesion images. As for the classification, the final layer of the Discriminator is labeled as a classifier for predicting the target class. This study deals with a binary classification predicting two classes-benign and malignant-in the ISIC2017 dataset: accuracy, recall, precision, and F1-score model classification performance. BAS measures classifier accuracy on imbalanced datasets. The DCGAN Classifier model demonstrated superior performance with a notable accuracy of 99.38% and 99% for recall, precision, F1 score, and BAS, outperforming the state-of-the-art deep learning models. These results show that the DCGAN Classifier can generate high-quality skin lesion images and accurately classify them, making it a promising tool for deep learning-based medical image analysis.

A New Artificial Intelligence Approach Using Extreme Learning Machine as the Potentially Effective Model to Predict and Analyze the Diagnosis of Anemia

The procedure to diagnose anemia is time-consuming and resource-intensive due to the existence of a multitude of symptoms that can be felt physically or seen visually. Anemia also has several forms, which can be distinguished based on several characteristics. It is possible to diagnose anemia through a quick, affordable, and easily accessible laboratory test known as the complete blood count (CBC), but the method cannot directly identify different kinds of anemia. Therefore, further tests are required to establish a gold standard for the type of anemia in a patient. These tests are uncommon in settings that offer healthcare on a smaller scale because they require expensive equipment. Moreover, it is also difficult to discern between beta thalassemia trait (BTT), iron deficiency anemia (IDA), hemoglobin E (HbE), and combination anemias despite the presence of multiple red blood cell (RBC) formulas and indices with differing optimal cutoff values. This is due to the existence of several varieties of anemia in individuals, making it difficult to distinguish between BTT, IDA, HbE, and combinations. Therefore, a more precise and automated prediction model is proposed to distinguish these four types to accelerate the identification process for doctors. Historical data were retrieved from the Laboratory of the Department of Clinical Pathology and Laboratory Medicine, Faculty of Medicine, Public Health, and Nursing, Universitas Gadjah Mada, Yogyakarta, Indonesia for this purpose. Furthermore, the model was developed using the algorithm for the extreme learning machine (ELM). This was followed by the measurement of the performance using the confusion matrix and 190 data representing the four classes, and the results showed 99.21% accuracy, 98.44% sensitivity, 99.30% precision, and an F1 score of 98.84%.

An Accelerated Maximally Split ADMM for a Class of Generalized Ridge Regression

Ridge regression (RR) has been commonly used in machine learning, but is facing computational challenges in big data applications. To meet the challenges, this article develops a highly parallel new algorithm, i.e., an accelerated maximally split alternating direction method of multipliers (A-MS-ADMM), for a class of generalized RR (GRR) that allows different regularization factors for different regression coefficients. Linear convergence of the new algorithm along with its convergence ratio is established. Optimal parameters of the algorithm for the GRR with a particular set of regularization factors are derived, and a selection scheme of the algorithm parameters for the GRR with general regularization factors is also discussed. The new algorithm is then applied in the training of single-layer feedforward neural networks. Experimental results on performance validation on real-world benchmark datasets for regression and classification and comparisons with existing methods demonstrate the fast convergence, low computational complexity, and high parallelism of the new algorithm.

Functional extreme learning machine for regression and classification

Although Extreme Learning Machine (ELM) can learn thousands of times faster than traditional slow gradient algorithms for training neural networks, ELM fitting accuracy is limited. This paper develops Functional Extreme Learning Machine (FELM), which is a novel regression and classifier. It takes functional neurons as the basic computing units and uses functional equation-solving theory to guide the modeling process of functional extreme learning machines. The functional neuron function of FELM is not fixed, and its learning process refers to the process of estimating or adjusting the coefficients. It follows the spirit of extreme learning and solves the generalized inverse of the hidden layer neuron output matrix through the principle of minimum error, without iterating to obtain the optimal hidden layer coefficients. To verify the performance of the proposed FELM, it is compared with ELM, OP-ELM, SVM and LSSVM on several synthetic datasets, XOR problem, benchmark regression and classification datasets. The experimental results show that although the proposed FELM has the same learning speed as ELM, its generalization performance and stability are better than ELM.

A study on the adsorption property and mechanism of β-cyclodextrin/polyvinyl alcohol/polyacrylic acid hydrogel for ciprofloxacin

Polyvinyl alcohol (PVA), acrylic acid (AA), and β-cyclodextrin (β-CD) were used as monomers, and ammonium persulfate was used as an initiator. Orthogonal tests were optimized the experimental condition, and aqueous polymerization was used to prepare poly-β-cyclodextrin/polyvinyl alcohol/polyacrylic acid (β-CD/PVA/PAA) hydrogel. The samples were characterized by FT-IR (Fourier transform infrared), SEM (Scanning electron microscopy), and XRD (X-ray diffraction). β-CD/PVA/PAA hydrogel was analyzed, which influenced external environmental factors on the β-CD/PVA/PAA hydrogel adsorption performance, and the kinetic behavior of β-CD/PVA/PAA hydrogel on ciprofloxacin (CIP) adsorption was explored. The results concluded that the prepared β-CD/PVA/PAA hydrogel has a well-defined three-dimensional network structure. The decrease in the pH of the CIP solution and the adsorption temperature reduces the adsorption reaction of β-CD/PVA/PAA hydrogel on CIP. The kinetics of CIP adsorption by β-CD/PVA/PAA hydrogel confirmed the pseudo-second-order kinetic model (R 2 > 0.997), the maximum equilibrium adsorption amounts is 372.12 mg/g, the removal rate reaches 74.42%. The adsorption process was mainly chemisorption, the adsorption isotherm fits the Freundlich adsorption isotherm model (R 2 > 0.946), and the adsorption process was heterogeneous with multi-molecular layer adsorption. The adsorption process inclined more toward the adsorption of inhomogeneous multi-molecular layers. The β-CD/PVA/PAA hydrogel retained 80% adsorption properties after three adsorption-desorption under optimal conditions.

Word2vec Word Embedding-Based Artificial Intelligence Model in the Triage of Patients with Suspected Diagnosis of Major Ischemic Stroke: A Feasibility Study

BACKGROUND

The possible benefits of using semantic language models in the early diagnosis of major ischemic stroke (MIS) based on artificial intelligence (AI) are still underestimated. The present study strives to assay the feasibility of the word2vec word embedding-based model in decreasing the risk of false negatives during the triage of patients with suspected MIS in the emergency department (ED).

METHODS

The main ICD-9 codes related to MIS were used for the 7-year retrospective data collection of patients managed at the ED with a suspected diagnosis of stroke. The data underwent "tokenization" and "lemmatization". The word2vec word-embedding algorithm was used for text data vectorization.

RESULTS

Out of 648 MIS, the word2vec algorithm successfully identified 83.9% of them, with an area under the curve of 93.1%.

CONCLUSIONS

Natural language processing (NLP)-based models in triage have the potential to improve the early detection of MIS and to actively support the clinical staff.

Artificial Intelligence in Spinal Imaging: Current Status and Future Directions

Spinal maladies are among the most common causes of pain and disability worldwide. Imaging represents an important diagnostic procedure in spinal care. Imaging investigations can provide information and insights that are not visible through ordinary visual inspection. Multiscale in vivo interrogation has the potential to improve the assessment and monitoring of pathologies thanks to the convergence of imaging, artificial intelligence (AI), and radiomic techniques. AI is revolutionizing computer vision, autonomous driving, natural language processing, and speech recognition. These revolutionary technologies are already impacting radiology, diagnostics, and other fields, where automated solutions can increase precision and reproducibility. In the first section of this narrative review, we provide a brief explanation of the many approaches currently being developed, with a particular emphasis on those employed in spinal imaging studies. The previously documented uses of AI for challenges involving spinal imaging, including imaging appropriateness and protocoling, image acquisition and reconstruction, image presentation, image interpretation, and quantitative image analysis, are then detailed. Finally, the future applications of AI to imaging of the spine are discussed. AI has the potential to significantly affect every step in spinal imaging. AI can make images of the spine more useful to patients and doctors by improving image quality, imaging efficiency, and diagnostic accuracy.

Multi-Class Skin Problem Classification Using Deep Generative Adversarial Network (DGAN)

The lack of annotated datasets makes the automatic detection of skin problems very difficult, which is also the case for most other medical applications. The outstanding results achieved by deep learning techniques in developing such applications have improved the diagnostic accuracy. Nevertheless, the performance of these models is heavily dependent on the volume of labelled data used for training, which is unfortunately not available. To address this problem, traditional data augmentation is usually adopted. Recently, the emergence of a generative adversarial network (GAN) seems a more plausible solution, where synthetic images are generated. In this work, we have developed a deep generative adversarial network (DGAN) multi-class classifier, which can generate skin problem images by learning the true data distribution from the available images. Unlike the usual two-class classifier, we have developed a multi-class solution, and to address the class-imbalanced dataset, we have taken images from different datasets available online. One main challenge faced during our development is mainly to improve the stability of the DGAN model during the training phase. To analyse the performance of GAN, we have developed two CNN models in parallel based on the architecture of ResNet50 and VGG16 by augmenting the training datasets using the traditional rotation, flipping, and scaling methods. We have used both labelled and unlabelled data for testing to test the models. DGAN has outperformed the conventional data augmentation by achieving a performance of 91.1% for the unlabelled dataset and 92.3% for the labelled dataset. On the contrary, CNN models with data augmentation have achieved a performance of up to 70.8% for the unlabelled dataset. The outcome of our DGAN confirms the ability of the model to learn from unlabelled datasets and yet produce a good diagnosis result.

Building Unmanned Store Identification Systems Using YOLOv4 and Siamese Network

Labor is the most expensive in retail stores. In order to increase the profit of retail stores, unmanned stores could be a solution for reducing labor cost. Deep learning is a good way for recognition, classification, and so on; in particular, it has high accuracy and can be implemented in real time. Based on deep learning, in this paper, we use multiple deep learning models to solve the problems often encountered in unmanned stores. Instead of using multiple different sensors, only five cameras are used as sensors to build a high-accuracy, low-cost unmanned store; for the full use of space, we then propose a method for calculating stacked goods, so that the space can be effectively used. For checkout, without a checking counter, we use a Siamese network combined with the deep learning model to directly identify products instantly purchased. As for protecting the store from theft, a new architecture was proposed, which can detect possible theft from any angle of the store and prevent unnecessary financial losses in unmanned stores. As all the customers’ buying records are identified and recorded in the server, it can be used to identify the popularity of the product. In particular, it can reduce the stock of unpopular products and reduce inventory.

A Neural Network Approach for Chinese Sports Tourism Demand Based on Knowledge Discovery

With the vigorous development of the Chinese economy and people's pursuit of quality, sports activities of people pursuit are no longer limited to simple physical exercise, but a way that pursues higher-quality sports tourism. As a new industry, it cannot guarantee that sports tourism will be accepted by all people, and it will be limited by geographical, economic, time, and other conditions. The participation number of Chinese sports tourism is more concerned by organizers or operators. Predicting the participation number of sports tourism based on the knowledge discovery method is meaningful and economical work. In this paper, a variety of sports tourism data are classified by clustering method, and the categories with similar characteristics are classified. Then, the convolution and long short-term memory hybrid neural network are used to extract the spatial and temporal information of sports tourism characteristics, which completes the prediction of Chinese sports tourism categories. The research results show that the clustering method has high accuracy for the classification of sports tourism categories, and the weights occupied by the categories are relatively uniform. The ConvLSTM neural network also has obvious advantages in predicting Chinese sports tourism methods. The largest error is only 2.89%, and the correlation coefficient also reaches 0.98, which is enough to be trusted for the prediction of Chinese sports tourism categories.

Real-Time Tracking of Object Melting Based on Enhanced DeepLab v3+ Network

In order to reveal the dissolution behavior of iron tailings in blast furnace slag, the main component of iron tailings, SiO2, was used for research. Aiming at the problem of information loss and inaccurate extraction of tracking molten SiO2 particles in high temperature, a method based on the improved DeepLab v3+ network was proposed to track, segment, and extract small object particles in real time. First, by improving the decoding layer of the DeepLab v3+ network, construct dense ASPP (atrous spatial pyramid pooling) modules with different dilation rates to optimize feature extraction, increase the shallow convolution of the backbone network, and merge it into the upper convolution decoding part to increase detailed capture. Secondly, integrate the lightweight network MobileNet v3 to reduce network parameters, further speed up image detection, and reduce the memory usage to achieve real-time image segmentation and adapt to low-level configuration hardware. Finally, improve the expression of the loss function for the binary classification model of small object in this paper, combining the advantages of the Dice Loss binary classification segmentation and the Focal Loss balance of positive and negative samples, solving the problem of unbalanced dataset caused by the small proportion of positive samples. Experimental results show that MIoU (mean intersection over union) of the proposed model for small object segmentation is 6% higher than that of the original model, the overall MIoU is increased by 3%, and the execution time and memory consumption are only half of the original model, which can be well applied to real-time tracking and segmentation of small particles.

Role of twitter user profile features in retweet prediction for big data streams

To study the various factors influencing the process of information sharing on Twitter is a very active research area. This paper aims to explore the impact of numerical features extracted from user profiles in retweet prediction from the real-time raw feed of tweets. The originality of this work comes from the fact that the proposed model is based on simple numerical features with the least computational complexity, which is a scalable solution for big data analysis. This research work proposes three new features from the tweet author profile to capture the unique behavioral pattern of the user, namely "Author total activity", "Author total activity per year", and "Author tweets per year". The features set is tested on a dataset of 100 million random tweets collected through Twitter API. The binary labels regression gave an accuracy of 0.98 for user-profile features and gave an accuracy of 0.99 when combined with tweet content features. The regression analysis to predict the retweet count gave an R-squared value of 0.98 with combined features. The multi-label classification gave an accuracy of 0.9 for combined features and 0.89 for user-profile features. The user profile features performed better than tweet content features and performed even better when combined. This model is suitable for near real-time analysis of live streaming data coming through Twitter API and provides a baseline pattern of user behavior based on numerical features available from user profiles only.

Ternary Compression for Communication-Efficient Federated Learning

Learning over massive data stored in different locations is essential in many real-world applications. However, sharing data is full of challenges due to the increasing demands of privacy and security with the growing use of smart mobile devices and Internet of thing (IoT) devices. Federated learning provides a potential solution to privacy-preserving and secure machine learning, by means of jointly training a global model without uploading data distributed on multiple devices to a central server. However, most existing work on federated learning adopts machine learning models with full-precision weights, and almost all these models contain a large number of redundant parameters that do not need to be transmitted to the server, consuming an excessive amount of communication costs. To address this issue, we propose a federated trained ternary quantization (FTTQ) algorithm, which optimizes the quantized networks on the clients through a self-learning quantization factor. Theoretical proofs of the convergence of quantization factors, unbiasedness of FTTQ, as well as a reduced weight divergence are given. On the basis of FTTQ, we propose a ternary federated averaging protocol (T-FedAvg) to reduce the upstream and downstream communication of federated learning systems. Empirical experiments are conducted to train widely used deep learning models on publicly available data sets, and our results demonstrate that the proposed T-FedAvg is effective in reducing communication costs and can even achieve slightly better performance on non-IID data in contrast to the canonical federated learning algorithms.

SUKRY: Suricata IDS with Enhanced kNN Algorithm on Raspberry Pi for Classifying IoT Botnet Attacks

The focus of this research is the application of the k-Nearest Neighbor algorithm in terms of classifying botnet attacks in the IoT environment. The kNN algorithm has several advantages in classification tasks, such as simplicity, effectiveness, and robustness. However, it does not perform well in handling large datasets such as the Bot-IoT dataset, which represents a huge amount of data about botnet attacks on IoT networks. Therefore, improving the kNN performance in classifying IoT botnet attacks is the main concern in this study by applying several feature selection techniques. The whole research process was conducted in the Rapidminer environment using three prebuilt feature selection techniques, namely, Information Gain, Forward Selection, and Backward Elimination. After comparing accuracy, precision, recall, F1 score and processing time, the combination of the kNN algorithm and the Forward Selection technique (kNN-FS) achieves the best results among others, with the highest level of accuracy and the fastest execution time among others. Finally, kNN-FS is used in developing SUKRY, which stands for Suricata IDS with Enhanced kNN Algorithm on Raspberry Pi.

Estimation of Salinity Content in Different Saline-Alkali Zones Based on Machine Learning Model Using FOD Pretreatment Method

Soil salinization is a global ecological and environmental problem in arid and semi-arid areas that can be ameliorated via soil management, visible-near infrared-shortwave infrared (VNIR-SWIR) spectroscopy can be adapted to rapidly monitor soil salinity content. This study explored the potential of Grünwald–Letnikov fractional-order derivative (FOD), feature band selection methods, nonlinear partial least squares regression (PLSR), and four machine learning models to estimate the soil salinity content using VNIR-SWIR spectra. Ninety sample points were field scanned with VNIR-SWR and soil samples (0–20 cm) were obtained at the time of scanning. The samples points come from three zones representing different intensities of human interference (I, II, and III Zones) in Fukang, Xinjiang, China. Each zone contained thirty sample points. For modeling, we firstly adopted FOD (with intervals of 0.1 and range of 0–2) as a preprocessing method to analyze soil hyperspectral data. Then, four sets of spectral bands (R-FOD-FULL indicates full band range, R-FOD-CC5 bands that met a 0.05 significance test, R-FOD-CC1 bands that met a 0.01 significance test, and R-FOD-CC1-CARS represents CC1 combined with competitive adaptive reweighted sampling) were selected as spectral input variables to develop the estimation model. Finally, four machine learning models, namely, generalized regression neural network (GRNN), extreme learning machine (ELM), random forest (RF), and PLSR, to estimate soil salinity. Study results showed that (1) the heat map of correlation coefficient matrix between hyperspectral data and salinity indicated that FOD significantly improved the correlation. (2) The characteristic band variables extracted and used by R-FOD-CC1 were fewer in number, and redundancy between bands smaller than R-FOD-FULL and R-FOD-CC5, thus estimation accuracy of R-FOD-CC1 was higher than R-FOD-CC5 or R-FOD-FULL. A high prediction accuracy was achieved with a less complex calculation. (3) The GRNN model yielded the best salinity estimation in all three zones compared to ELM, BPNN, RF, and PLSR on the whole, whereas, the RF model had the worst estimation effect. The R-FOD-CC1-CARS-GRNN model yielded the best salinity estimation in I Zone with R2, RMSE and RPD of 0.7784, 1.8762, and 2.0568, respectively. The fractional order was 1.5 and estimation performance was great. The optimal model for predicting soil salinity in II and III Zone was, also, R-FOD-CC1-CARS-GRNN (R2 = 0.7912, RMSE = 3.4001, and RPD = 1.8985 in II Zone; R2 = 0.8192, RMSE = 6.6260, and RPD = 1.8190 in III Zone), with the fractional order of 1.7- and 1.6-, respectively, and the estimation performance were all fine. (4) The characteristic bands selected by the best model in I, II, and III Zones were 8, 9, and 11, respectively, which account for 0.45%, 0.51%, and 0.63%% of the full bands. This approach reduces the number of modeled band variables and simplifies the model structure.

A Parallel Algorithm for Dividing Octonions

The article presents a parallel hardware-oriented algorithm designed to speed up the division of two octonions. The advantage of the proposed algorithm is that the number of real multiplications is halved as compared to the naive method for implementing this operation. In the synthesis of the discussed algorithm, the matrix representation of this operation was used, which allows us to present the division of octonions by means of a vector–matrix product. Taking into account a specific structure of the matrix multiplicand allows for reducing the number of real multiplications necessary for the execution of the octonion division procedure.

Using Big Data-Based Neural Network Parallel Optimization Algorithm in Sports Fatigue Warning

In high-paced and efficient life and work, fatigue is one of the important factors that cause accidents such as traffic and medical accidents. This study designs a feature map-based pruning strategy (PFM), which effectively reduces redundant parameters and reduces the time and space complexity of parallelized deep convolutional neural network (DCNN) training; a correction is proposed in the Map stage. The secant conjugate gradient method (CGMSE) realizes the fast convergence of the conjugate gradient method and improves the convergence speed of the network; in the Reduce stage, a load balancing strategy to control the load rate (LBRLA) is proposed to achieve fast and uniform data grouping to ensure the parallelization performance of the parallel system. Finally, the related fatigue algorithm's research and simulation based on the human eye are carried out on the PC. The human face and eye area are detected from the video image collected using the USB camera, and the frame difference method and the position information of the human eye on the face are used. To track the human eye area, extract the relevant human eye fatigue characteristics, combine the blink frequency, closed eye duration, PERCLOS, and other human eye fatigue determination mechanisms to determine the fatigue state, and test and verify the designed platform and algorithm through experiments. This system is designed to enable people who doze off, such as drivers, to discover their state in time through the system and reduce the possibility of accidents due to fatigue.

Supervised Feature Selection With Orthogonal Regression and Feature Weighting

Effective features can improve the performance of a model and help us understand the characteristics and underlying structure of complex data. Previously proposed feature selection methods usually cannot retain more discriminative information. To address this shortcoming, we propose a novel supervised orthogonal least square regression model with feature weighting for feature selection. The optimization problem of the objective function can be solved by employing generalized power iteration and augmented Lagrangian multiplier methods. Experimental results show that the proposed method can more effectively reduce feature dimensionality and obtain better classification results than traditional feature selection methods. The convergence of our iterative method is also proved. Consequently, the effectiveness and superiority of the proposed method are verified both theoretically and experimentally.

Few-shot pulse wave contour classification based on multi-scale feature extraction

The annotation procedure of pulse wave contour (PWC) is expensive and time-consuming, thereby hindering the formation of large-scale datasets to match the requirements of deep learning. To obtain better results under the condition of few-shot PWC, a small-parameter unit structure and a multi-scale feature-extraction model are proposed. In the small-parameter unit structure, information of adjacent cells is transmitted through state variables. Simultaneously, a forgetting gate is used to update the information and retain long-term dependence of PWC in the form of unit series. The multi-scale feature-extraction model is an integrated model containing three parts. Convolution neural networks are used to extract spatial features of single-period PWC and rhythm features of multi-period PWC. Recursive neural networks are used to retain the long-term dependence features of PWC. Finally, an inference layer is used for classification through extracted features. Classification experiments of cardiovascular diseases are performed on photoplethysmography dataset and continuous non-invasive blood pressure dataset. Results show that the classification accuracy of the multi-scale feature-extraction model on the two datasets respectively can reach 80% and 96%, respectively.

Recursive Block LU Decomposition based ELM in Apache Spark

Due to the massive memory and computational resources required to build complex machine learning models on large datasets, many researchers are employing distributed environments for training the models on large datasets. The parallel implementations of Extreme Learning Machine (ELM) with many variants have been developed using MapReduce and Spark frameworks in the recent years. However, these approaches have severe limitations in terms of Input-Output (I/O) cost, memory, etc. From the literature, it is known that the complexity of ELM is directly propositional to the computation of Moore-Penrose pseudo inverse of hidden layer matrix in ELM. Most of the ELM variants developed on Spark framework have employed Singular Value Decomposition (SVD) to compute the Moore-Penrose pseudo inverse. But, SVD has severe memory limitations when experimenting with large datasets. In this paper, a method that uses Recursive Block LU Decomposition to compute the Moore-Penrose generalized inverse over the Spark cluster has been proposed to reduce the computational complexity. This method enhances the ELM algorithm to be efficient in handling the scalability and also having faster execution of the model. The experimental results have shown that the proposed method is efficient than the existing algorithms available in the literature.

LiDAR Point Cloud Recognition and Visualization with Deep Learning for Overhead Contact Inspection

As overhead contact (OC) is an essential part of power supply systems in high-speed railways, it is necessary to regularly inspect and repair abnormal OC components. Relative to manual inspection, applying LiDAR (light detection and ranging) to OC inspection can improve efficiency, accuracy, and safety, but it faces challenges to efficiently and effectively segment LiDAR point cloud data and identify catenary components. Recent deep learning-based recognition methods are rarely employed to recognize OC components, because they have high computational complexity, while their accuracy needs to be improved. To track these problems, we first propose a lightweight model, RobotNet, with depthwise and pointwise convolutions and an attention module to recognize the point cloud. Second, we optimize RobotNet to accelerate its recognition speed on embedded devices using an existing compilation tool. Third, we design software to facilitate the visualization of point cloud data. Our software can not only display a large amount of point cloud data, but also visualize the details of OC components. Extensive experiments demonstrate that RobotNet recognizes OC components more accurately and efficiently than others. The inference speed of the optimized RobotNet increases by an order of magnitude. RobotNet has lower computational complexity than other studies. The visualization results also show that our recognition method is effective.

A Maximally Split and Relaxed ADMM for Regularized Extreme Learning Machines

One of the salient features of the extreme learning machine (ELM) is its fast learning speed. However, in a big data environment, the ELM still suffers from an overly heavy computational load due to the high dimensionality and the large amount of data. Using the alternating direction method of multipliers (ADMM), a convex model fitting problem can be split into a set of concurrently executable subproblems, each with just a subset of model coefficients. By maximally splitting across the coefficients and incorporating a novel relaxation technique, a maximally split and relaxed ADMM (MS-RADMM), along with a scalarwise implementation, is developed for the regularized ELM (RELM). The convergence conditions and the convergence rate of the MS-RADMM are established, which exhibits linear convergence with a smaller convergence ratio than the unrelaxed maximally split ADMM. The optimal parameter values of the MS-RADMM are obtained and a fast parameter selection scheme is provided. Experiments on ten benchmark classification data sets are conducted, the results of which demonstrate the fast convergence and parallelism of the MS-RADMM. Complexity comparisons with the matrix-inversion-based method in terms of the numbers of multiplication and addition operations, the computation time and the number of memory cells are provided for performance evaluation of the MS-RADMM.

Stability-Based Generalization Analysis of Distributed Learning Algorithms for Big Data

As one of the efficient approaches to deal with big data, divide-and-conquer distributed algorithms, such as the distributed kernel regression, bootstrap, structured perception training algorithms, and so on, are proposed and broadly used in learning systems. Some learning theories have been built to analyze the feasibility, approximation, and convergence bounds of these distributed learning algorithms. However, less work has been studied on the stability of these distributed learning algorithms. In this paper, we discuss the generalization bounds of distributed learning algorithms from the view of algorithmic stability. First, we introduce a definition of uniform distributed stability for distributed algorithms and study the distributed algorithms' generalization risk bounds. Then, we analyze the stability properties and generalization risk bounds of a kind of regularization-based distributed algorithms. Two generalization distributed risks obtained show that the generalization distributed risk bounds for the difference between their generalization distributed and empirical distributed/leave-one-computer-out risks are closely related to the size of samples n and the amount of working computers m as O(m/n^1/2) . Furthermore, the results in this paper indicate that, for a good generalization regularized distributed kernel algorithm, the regularization parameter λ should be adjusted with the change of the term m/n^1/2 . These theoretic discoveries provide the useful guidance when deploying the distributed algorithms on practical big data platforms. We explore our theoretic analyses through two simulation experiments. Finally, we discuss some problems about the sufficient amount of working computers, nonequivalence, and generalization for distributed learning. We show that the rules for the computation on one single computer may not always hold for distributed learning.

A novel molecular representation with BiGRU neural networks for learning atom

Molecular representations play critical roles in researching drug design and properties, and effective methods are beneficial to assisting in the calculation of molecules and solving related problem in drug discovery. In previous years, most of the traditional molecular representations are based on hand-crafted features and rely heavily on biological experimentations, which are often costly and time consuming. However, recent researches achieve promising results using machine learning on various domains. In this article, we present a novel method named Smi2Vec-BiGRU that is designed for learning atoms and solving the single- and multitask binary classification problems in the field of drug discovery, which are the basic and also key problems in this field. Specifically, our approach transforms the molecule data in the SMILES format into a set of sample vectors and then feeds them into the bidirectional gated recurrent unit neural networks for training, which learns low-dimensional vector representations for molecular drug. We conduct extensive experiments on several widely used benchmarks including Tox21, SIDER and ClinTox. The experimental results show that our approach can achieve state-of-the-art performance on these benchmarking datasets, demonstrating the feasibility and competitiveness of our proposed approach.

Automatically Detecting Excavator Anomalies Based on Machine Learning

Excavators are one of the most frequently used pieces of equipment in large-scale construction projects. They are closely related to the construction speed and total cost of the entire project. Therefore, it is very important to effectively monitor their operating status and detect abnormal conditions. Previous research work was mainly based on expert systems and traditional statistical models to detect excavator anomalies. However, these methods are not particularly suitable for modern sophisticated excavators. In this paper, we take the first step and explore the use of machine learning methods to automatically detect excavator anomalies by mining its working condition data collected from multiple sensors. The excavators we studied are from Sany Group, the largest construction machinery manufacturer in China. We have collected 40 days working condition data of 107 excavators from Sany. In addition, we worked with six excavator operators and engineers for more than a month to clean the original data and mark the anomalous samples. Based on the processed data, we have designed three anomaly detection schemes based on machine learning methods, using support vector machine (SVM), back propagation (BP) neural network and decision tree algorithms, respectively. Based on the real excavator data, we have carried out a comprehensive evaluation. The results show that the anomaly detection accuracy is as high as 99.88%, which is obviously superior to the previous methods based on expert systems and traditional statistical models.

Extreme Learning Machine With Affine Transformation Inputs in an Activation Function

The extreme learning machine (ELM) has attracted much attention over the past decade due to its fast learning speed and convincing generalization performance. However, there still remains a practical issue to be approached when applying the ELM: the randomly generated hidden node parameters without tuning can lead to the hidden node outputs being nonuniformly distributed, thus giving rise to poor generalization performance. To address this deficiency, a novel activation function with an affine transformation (AT) on its input is introduced into the ELM, which leads to an improved ELM algorithm that is referred to as an AT-ELM in this paper. The scaling and translation parameters of the AT activation function are computed based on the maximum entropy principle in such a way that the hidden layer outputs approximately obey a uniform distribution. Application of the AT-ELM algorithm in nonlinear function regression shows its robustness to the range scaling of the network inputs. Experiments on nonlinear function regression, real-world data set classification, and benchmark image recognition demonstrate better performance for the AT-ELM compared with the original ELM, the regularized ELM, and the kernel ELM. Recognition results on benchmark image data sets also reveal that the AT-ELM outperforms several other state-of-the-art algorithms in general.

Shared Nearest-Neighbor Quantum Game-Based Attribute Reduction With Hierarchical Coevolutionary Spark and Its Application in Consistent Segmentation of Neonatal Cerebral Cortical Surfaces

The unprecedented increase in data volume has become a severe challenge for conventional patterns of data mining and learning systems tasked with handling big data. The recently introduced Spark platform is a new processing method for big data analysis and related learning systems, which has attracted increasing attention from both the scientific community and industry. In this paper, we propose a shared nearest-neighbor quantum game-based attribute reduction (SNNQGAR) algorithm that incorporates the hierarchical coevolutionary Spark model. We first present a shared coevolutionary nearest-neighbor hierarchy with self-evolving compensation that considers the features of nearest-neighborhood attribute subsets and calculates the similarity between attribute subsets according to the shared neighbor information of attribute sample points. We then present a novel attribute weight tensor model to generate ranking vectors of attributes and apply them to balance the relative contributions of different neighborhood attribute subsets. To optimize the model, we propose an embedded quantum equilibrium game paradigm (QEGP) to ensure that noisy attributes do not degrade the big data reduction results. A combination of the hierarchical coevolutionary Spark model and an improved MapReduce framework is then constructed that it can better parallelize the SNNQGAR to efficiently determine the preferred reduction solutions of the distributed attribute subsets. The experimental comparisons demonstrate the superior performance of the SNNQGAR, which outperforms most of the state-of-the-art attribute reduction algorithms. Moreover, the results indicate that the SNNQGAR can be successfully applied to segment overlapping and interdependent fuzzy cerebral tissues, and it exhibits a stable and consistent segmentation performance for neonatal cerebral cortical surfaces.