Search

Bitter taste receptors were recently found to be involved in numerous physiological and pathological conditions other than taste and are suggested as potential drug targets. In vivo and in vitro techniques for screening bitterants as ligands come with economical, time and ethic challenges. Therefore, in silico tools can represent a valuable alternative due to their practicality. Yet, the main challenge of already established ligand-based (LB) classifiers is the low number of experimentally confirmed bitterants and non-bitterants. Premexotac models were constructed as a LB bitterants screener, exploring novel combinations of feature extraction, feature selection and learning algorithms as a contrast with the already available screeners. Premexotac came among the top performers, exhibiting a F-1 score up to 81% on external validation. Premexotac identified as well insights on physicochemical and topological descriptors important for bitter prediction. Among the key insights, important molecular substructures from Extended Connectivity Fingerprints for bitterness classification were identified. Also, the importance of a selection of physicochemical/topological descriptors was ranked using mutual information and it was found that descriptors related to the ramification of the molecular structure and molecular weight came at the top of the ranking. The remaining challenges for improving performance were discussed and stated, widening the LB bitterness prediction outlook., Competing Interests: Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (Copyright © 2022 Elsevier B.V. All rights reserved.)

Background: Health-related quality of life (HRQoL) and well-being refer to the positive, subjective state that is contrary to illness. HRQoL instruments include some common questionnaires, which may often be understood differently depending on the level of individuals' knowledge., Aims: To investigate the ability of 36 Short Form Health Survey (SF-36) as a well-known questionnaire in evaluating people's well-being., Methods: We compared unsupervised artificial neural networks with a self-organized map learning algorithm and k-means clustering method. Understanding of the content of the questionnaire was also checked according to age group and sex. The study included 1087 people aged > 18 years (640 healthy individuals and 447 patients with chronic diseases) in Shiraz, Islamic Republic of Iran between 2011 and 2013., Results: The eight subscale scores of the SF-36 instrument were not able to evaluate the well-being of people. The ability of all 36 items in the questionnaire was > 60% in both self-organized map and k-means methods. The self-organized map learning algorithm evaluated people better than the k-means clustering method, based on the accuracy rate in prediction. The SF-36 instrument was better understood by young people., Conclusions: Differences in people's health conditions may not appear on the SF-36 subscale scores; therefore, the findings from the subscale scores of SF-36 should be cautiously interpreted., (Copyright © World Health Organization (WHO) 2019. Some rights reserved. This work is available under the CC BY-NC-SA 3.0 IGO license (https://creativecommons.org/licenses/by-nc-sa/3.0/igo).)

Optimized scheduling is an important task in the scheduling of job execution on cloud virtual machines, where optimal resource usage and a shorter makespan have become important features. When scheduling jobs, balancing the workload across all available virtual machines provides optimized performance. Reinforcement learning is a better optimization algorithm due to its adaptability to dynamic environments and balancing exploration and exploitation. To perform optimized balancing of job scheduling, an Actor- Critic-based reinforcement algorithm is applied in this work. The Alibaba cloud dataset is used to analyze the algorithm's performance. Policy constraints are made for assigning the number of tasks to the scheduler. During the learning phase, the rewards turn out to be negative. After the learning phase, the rewards stabilize. The results show that the algorithm is able to produce positive reward points. A 5% reduction in the makespan of job execution demonstrates the improvement in scheduling and resource use. [ABSTRACT FROM AUTHOR]

Copyright of Remote Sensing for Natural Resources is the property of Remote Sensing for Natural Resources Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This paper addresses the challenges of traditional spindle assembly methods, which rely on trial-and-error approaches, hindering new product development evaluation. A stacking ensemble model is proposed to predict the assembly quality variation of machine tool spindles. The model uses data from 925 single-spindle inspections and extracts evaluation metrics from multiple domains to extract valuable information. Feature selection is performed using a correlation model to identify important features, and various lightweight supervised learning algorithms are applied to analyze the data. To further enhance the model's performance, a stacking ensemble approach is proposed, which combines algorithms. The results demonstrate that the proposed stacking ensemble model is an effective approach for predicting the assembly quality variation of machine tool spindles, using the data available. The proposed ensemble model enhances quality control processes in spindle assembly, enabling practitioners to identify key features and predict machine tool spindle assembly quality variations more accurately. [ABSTRACT FROM AUTHOR]

Efficient Learning Algorithms for the Best Capped Base-Stock Policy in Lost Sales Inventory Systems Periodic review, lost sales inventory systems with lead times are notoriously challenging to optimize. Recently, the capped base-stock policy, which places orders to bring the inventory position up to the order-up-to level subject to the order cap, has demonstrated exceptional performance. In the paper "UCB-Type Learning Algorithms with Kaplan–Meier Estimator for Lost Sales Inventory Models with Lead Times," Lyu, Zhang, and Xin propose an upper confidence bound–type learning framework. This framework, which incorporates simulations with the Kaplan–Meier estimator, works with censored demand observations. It can be applied to determine the optimal capped base-stock policy with a tight regret with respect to the planning horizon and the optimal base-stock policy with a regret that matches the best existing result. Both theoretical analysis and extensive numerical experiments demonstrate the effectiveness of the proposed learning framework. In this paper, we consider a classic periodic-review lost-sales inventory system with lead times, which is notoriously challenging to optimize with a wide range of real-world applications. We consider a joint learning and optimization problem in which the decision maker does not know the demand distribution a priori and can only use past sales information (i.e., censored demand). Departing from existing learning algorithms on this learning problem that require the convexity property of the underlying system, we develop an upper confidence bound (UCB)-type learning framework that incorporates simulations with the Kaplan–Meier estimator and demonstrate its applicability to learning not only the optimal capped base-stock policy in which convexity no longer holds, but also the optimal base-stock policy with a regret that matches the best existing result. Compared with a classic multi-armed bandit problem, our problem has unique challenges because of the nature of the inventory system, because (1) each action has long-term impacts on future costs, and (2) the system state space is exponentially large in the lead time. As such, our learning algorithms are not naive adoptions of the classic UCB algorithm; in fact, the design of the simulation steps with the Kaplan–Meier estimator and averaging steps is novel in our algorithms, and the confidence width in the UCB index is also different from the classic one. We prove the regrets of our learning algorithms are tight up to a logarithmic term in the planning horizon T. Our extensive numerical experiments suggest the proposed algorithms (almost) dominate existing learning algorithms. We also demonstrate how to select which learning algorithm to use with limited demand data. Supplemental Material: The online appendix is available at https://doi.org/10.1287/opre.2022.0273. [ABSTRACT FROM AUTHOR]

In the manufacturing industry, production efficiency is the core competitiveness of enterprises, but also an important goal of enterprise production planning management, and in recent years, the production efficiency of China's manufacturing industry does not meet the requirements, and there are still many problems to be solved. In the face of this situation, the relevant fields have explored intelligent manufacturing and proposed intelligent manufacturing technology, which mainly integrates information technology with traditional manufacturing system, can realize the important measures of intelligence, automation, digitalization and information in the manufacturing process, and can help solve the current key problems. Therefore, this paper takes ship manufacturing as the background. The real-time optimization technology of intelligent ship manufacturing based on learning algorithm is studied. This technology can effectively improve ship production efficiency and reduce production cost. [ABSTRACT FROM AUTHOR]

The long-time test of battery discharge time is the main problem for implementing the performance requirements of the standard IEC 60086. Prediction of battery discharge time can be a solution for the long-time test problem. The discharge test data conceive valuable information of the nonlinear characteristics of the battery to build battery models. An intelligent system with learning algorithm to update the capability of the model from a set of training test data is suitable to predict the battery discharge time. The well-known ANFIS (Adaptive Neuro Fuzzy Inference Systems) is a neuro-adaptive system that has adaptive input nodes and has been applied to many prediction problems through learning and validation of the network from data. In this paper, the empirical test data related to the discharge time are acquired from the laboratory tests and are used to build the persuasive ANFIS model for the discharge time prediction from inputs to 48 recombination output data by means of modified membership functions. [ABSTRACT FROM AUTHOR]

Artificial neural networks (ANNs) have led to landmark changes in many fields, but they still differ significantly from the mechanisms of real biological neural networks and face problems such as high computing costs, excessive computing power, and so on. Spiking neural networks (SNNs) provide a new approach combined with brain-like science to improve the computational energy efficiency, computational architecture, and biological credibility of current deep learning applications. In the early stage of development, its poor performance hindered the application of SNNs in real-world scenarios. In recent years, SNNs have made great progress in computational performance and practicability compared with the earlier research results, and are continuously producing significant results. Although there are already many pieces of literature on SNNs, there is still a lack of comprehensive review on SNNs from the perspective of improving performance and practicality as well as incorporating the latest research results. Starting from this issue, this paper elaborates on SNNs along the complete usage process of SNNs including network construction, data processing, model training, development, and deployment, aiming to provide more comprehensive and practical guidance to promote the development of SNNs. Therefore, the connotation and development status of SNN computing is reviewed systematically and comprehensively from four aspects: composition structure, data set, learning algorithm, software/hardware development platform. Then the development characteristics of SNNs in intelligent computing are summarized, the current challenges of SNNs are discussed and the future development directions are also prospected. Our research shows that in the fields of machine learning and intelligent computing, SNNs have comparable network scale and performance to ANNs and the ability to challenge large datasets and a variety of tasks. The advantages of SNNs over ANNs in terms of energy efficiency and spatial-temporal data processing have been more fully exploited. And the development of programming and deployment tools has lowered the threshold for the use of SNNs. SNNs show a broad development prospect for brain-like computing. [ABSTRACT FROM AUTHOR]

In this paper, different learning methods based on Artificial Neural Networks (ANNs) are examined to replace the default speed controller for high-precision position control and drift attenuation in robotic manipulators. ANN learning methods including Levenberg–Marquardt and Bayesian Regression are implemented and compared using a UR5 robot with six degrees of freedom to improve trajectory tracking and minimize position error. Extensive simulation and experimental tests on the identification and control of the robot by means of the neural network controllers yield comparable results with respect to the classical controller, showing the feasibility of the proposed approach. [ABSTRACT FROM AUTHOR]

This ongoing work is vehemently dedicated to the investigation of a class of ordinary linear Volterra type integro-differential equations with fractional order in numerical mode. By replacing the unknown function by an appropriate multilayered feed-forward type neural structure, the fractional problem of such initial value is changed into a course of non-linear minimization equations, to some extent. Put differently, interest was sparked in structuring an optimized iterative first-order algorithm to estimate solutions for the origin fractional problem. On top of that, some computer simulation models exemplify the preciseness and well-functioning of the indicated iterative technique. The outstanding accomplished numerical outcomes conveniently reflect the productivity and competency of artificial neural network methods compared to customary approaches. [ABSTRACT FROM AUTHOR]

This paper presents a method for solving linear Fredholm integro-differential equations using a feedforward neural network based on Legendre polynomials. Firstly, the Legendre polynomials are used to approximate the unknown function and the kernel function in the equations. Secondly, The roots of Legendre polynomials are obtained by using Newton iteration method to obtain Gaussian integration points, and the obtained Gaussian integration points are used as the input nodes of the neural network, and the corresponding weight is learned by the gradient descent method to obtain an approximate solution. Finally, through the numerical value Case analysis to verify the effectiveness of the method. [ABSTRACT FROM AUTHOR]

The conventional tracking-learning-detection (TLD) algorithm is sensitive to illumination changes, clutter, significant changes of target shape between consecutive frames. In addition, low frame rate scenarios result in drift in object position or even missing the object. To solve these problems and enhance the tracking robustness, in this paper, TLD algorithm was extended in two folds. First, Kalman filter and Mean-shift algorithms were combined for the tracking part and second, co-training semi-supervised learning algorithm was used for the learning part of the conventional TLD structure. The Kalman filter estimates the position of the target in the next frame based on the previous positions of the target. This reduces tracking failure. On the other hands, the Mean-shift tracking algorithm is robust to rotation, partial occlusion and scale changing. In the learning part of TLD structure, two training tracking algorithms with two independent classifiers were run on the current frame simultaneously. Its structure makes data of both pools (color features and target templates) update by the results of other algorithm, in addition to the results of the corresponding algorithm in each of the tracking and detection algorithms. Therefore, classifiers can learn faster changing features of the target during the consecutive frames in online tracking process. Finally, the extended structure can solve the problem of lost object in LFR videos tracking and other similar challenges simultaneously. In terms of overlap ratio metric, comparing with conventional TLD and extended kernelized correlation filters (EKCF) algorithms, the success rate of our algorithm under various scenarios has increased by 161.03% and 255.82% respectively and under other scenarios, in terms of precision metric, it has increased by 18.46% and 1479.47%, respectively. Accordingly, comparative evaluations of the proposed method to other top state-of-the-art tracking algorithms under various scenarios present in TB-100 dataset, the superior performance of the proposed algorithm was demonstrated compared to other tracking algorithms in terms of tracking robustness, stability and real-time performance. [ABSTRACT FROM AUTHOR]

The use of quantitative techniques for the classification of population segments is a critical phase to evaluate their conditions. This information will serve as input for planning strategies to alleviate poverty. In this article, we present a model of vector support machines. Consequently, a sample of families residing in Cartagena de Indias is segmented, based on certain economic and sociodemographic variables. Analytical results confirm that most important factors are employment status, accessibility to public services and familiar income. In addition, it is corroborated that neighborhood conditions and monetary transfers have a low discriminatory power. [ABSTRACT FROM AUTHOR]

Advances in technology have transformed and expanded the ways in which policing is run. One new manifestation is the mass acquisition and processing of private facial images via automatic facial recognition by the police: what we conceptualise as AFR-based policing. However, there is still a lack of clarity on the manner and extent to which this largely-unregulated technology is used by law enforcement agencies and on its impact on fundamental rights. Social understanding and involvement are still insufficient in the context of AFR technologies, which in turn affects social trust in and legitimacy and effectiveness of intelligent governance. This article delineates the function creep of this new concept, identifying the individual and collective harms it engenders. A technological, contextual perspective of the function creep of AFR in policing will evidence the comprehensive creep of training datasets and learning algorithms, which have by-passed an ignorant public. We thus argue individual harms to dignity, privacy and autonomy, combine to constitute a form of cultural harm, impacting directly on individuals and society as a whole. While recognising the limitations of what the law can achieve, we conclude by considering options for redress and the creation of an enhanced regulatory and oversight framework model, or Code of Conduct, as a means of encouraging cultural change from prevailing police indifference to enforcing respect for the human rights violations potentially engaged. The imperative will be to strengthen the top-level design and technical support of AFR policing, imbuing it with the values implicit in the rule of law, democratisation and scientisation-to enhance public confidence and trust in AFR social governance, and to promote civilised social governance in AFR policing. [ABSTRACT FROM AUTHOR]

Copyright of Journal of South China University of Technology (Natural Science Edition) is the property of South China University of Technology and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Introduction: Spiking Neural Networks (SNNs) are gaining significant traction in machine learning tasks where energy-efficiency is of utmost importance. Training such networks using the state-of-the-art back-propagation through time (BPTT) is, however, very time-consuming. Previous work employs an efficient GPU-accelerated backpropagation algorithm called SLAYER, which speeds up training considerably. SLAYER, however, does not take into account the neuron reset mechanism while computing the gradients, which we argue to be the source of numerical instability. To counteract this, SLAYER introduces a gradient scale hyper parameter across layers, which needs manual tuning. Methods: In this paper, we modify SLAYER and design an algorithm called EXODUS, that accounts for the neuron reset mechanism and applies the Implicit Function Theorem (IFT) to calculate the correct gradients (equivalent to those computed by BPTT). We furthermore eliminate the need for ad-hoc scaling of gradients, thus, reducing the training complexity tremendously. Results: We demonstrate, via computer simulations, that EXODUS is numerically stable and achieves comparable or better performance than SLAYER especially in various tasks with SNNs that rely on temporal features. [ABSTRACT FROM AUTHOR]

The conventional dermatology practice of performing noninvasive screening tests to detect skin diseases is a source of escapable diagnostic inaccuracies. Literature suggests that automated diagnosis is essential for improving diagnostic accuracies in medical fields such as dermatology, mammography, and colonography. Classification is an essential component of an assisted automation process that is rapidly gaining attention in the discipline of artificial intelligence for successful diagnosis, treatment, and recovery of patients. However, classifying skin lesions into multiple classes is challenging for most machine learning algorithms, especially for extremely imbalanced training datasets. This study proposes a novel ensemble deep learning algorithm based on the residual network with the next dimension and the dual path network with confidence preservation to improve the classification performance of skin lesions. The distributed computing paradigm was applied in the proposed algorithm to speed up the inference process by a factor of 0.25 for a faster classification of skin lesions. The algorithm was experimentally compared with 16 deep learning and 12 ensemble deep learning algorithms to establish its discriminating prowess. The experimental comparison was based on dermoscopic images congregated from the publicly available international skin imaging collaboration databases. We propitiously recorded up to 82.52% average sensitivity, 99.00% average specificity, 98.54% average balanced accuracy, and 92.84% multiclass accuracy without prior segmentation of skin lesions to outstrip numerous state-of-the-art deep learning algorithms investigated. [ABSTRACT FROM AUTHOR]

Chaos dynamics is an interesting nonlinear effect that can be observed in many chemical, electrical, and mechanical systems. The chaos phenomenon has many applications in various branches of engineering. On the other hand, the control of mobile robots to track unpredictable chaotic trajectories has a valuable application in many security problems and military missions. The main objective in this problem is to design a controller such that the robot tracks a desired chaotic path. In this paper, the concept of synchronization of chaotic systems is studied, and a new type-3 fuzzy system (T3FLS)-based controller is designed. The T3FLS is learned by some new adaptive rules. The new learning scheme of T3FLS helps to better stabilize and synchronize. The suggested controller has a better ability to cope with high-level uncertainties. Because, in addition to the fact that the T3FLSs have better ability in an uncertain environment, the designed compensator also improves the accuracy and robustness. Several simulations show better synchronization and control accuracy of the designed controller. [ABSTRACT FROM AUTHOR]

The size of datasets is growing exponentially as information technology advances, and it is becoming more and more crucial to provide efficient learning algorithms for neural networks to handle massive amounts of data. Due to their potential for handling huge datasets, feed-forward neural networks with random weights (FNNRWs) have drawn a lot of attention. In this paper, we introduced an efficient feed-forward neural network scheme (FNNS) for processing massive datasets with random weights. The FNNS divides large-scale data into subsets of the same size, and each subset derives the corresponding submodel. According to the activation function, the optimal range of input weights and biases is calculated. The input weight and biases are randomly generated in this range, and the iterative scheme is used to evaluate the output weight. The MNIST dataset was used as the basis for experiments. The experimental results demonstrate that the algorithm has a promising future in processing massive datasets. [ABSTRACT FROM AUTHOR]

The classification of high dimensional gene expression data is key to the development of effective diagnostic and prognostic tools. Feature selection involves finding the best subset with the highest power in predicting class labels. Here, we conducted a comparative study focused on different combinations of feature selectors (Chi-Squared, mRMR, Relief-F, and Genetic Algorithms) and classification learning algorithms (Random Forests, PLS-DA, SVM, Regularized Logistic/Multinomial Regression, and kNN) to identify those with the best predictive capacity. The performance of each combination is evaluated through an empirical study on three benchmark cancer-related microarray datasets. Our results first suggest that the quality of the data relevant to the target classes is key for the successful classification of cancer phenotypes. We also proved that, for a given classification learning algorithm and dataset, all filters have a similar performance. Interestingly, filters achieve comparable or even better results with respect to the GA-based wrappers, while also being easier and faster to implement. Taken together, our findings suggest that simple, well-established feature selectors in combination with optimized classifiers guarantee good performances, with no need for complicated and computationally demanding methodologies. [ABSTRACT FROM AUTHOR]

Copyright of Industrial Minerals & Processing / Huagong Kuangwu yu Jiagong is the property of Industrial Minerals & Processing Editorial Office and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This article deals with the development of four modified radial basis function neural network (RBFNN) models. The corresponding learning algorithms associated with the updating of internal parameters of the models are derived. The conventional inputs are used in the first and second modified RBFNN models (models 3 and 4) whereas exponential nonlinear inputs are used in the fifth and sixth RBFNN models to provide additional nonlinearity for achieving a better solution of nonlinear classification, and direct and inverse modeling problems. To assess and compare the performance potentiality of the proposed four new RBFNN models, one classification problem, one direct modeling problem, and one inverse modeling problem are solved through computer simulation-based experiments. For comparison and to assign the performance rank of each of the four modified RBFNN models, two conventional and commonly used RBFNN models (models 1 and 2) are also simulated. To access the performance of different models during the training phase of Examples 1 and 2, the root mean-square error (RMSE) value, mean absolute deviation (MAD), and the number of iterations required to achieve convergence are obtained. For the third example, only the first two performance measures are found. During the testing or validation phase, the output responses of the different models of Example 2 are compared with the desired response analysis. For Example 3, the bit-error rate (BER) plots are compared. The observation of all the results demonstrates consistent ranks of all models in the case of all three examples. It is, in general, found that the ranks of the models 1–6 are 6, 4, 3, 2, 5, and 1, respectively. In essence, in terms of all performance measures, model M-6 with an exponential version of inputs with weights on both layers occupies the first position whereas model M-4 with conventional inputs as the second position. [ABSTRACT FROM AUTHOR]

In this paper, a multimodal biometric identification system based on feature level fusion and machine learning techniques is described. The significance of this study relates to the combination of face and palm print for an individual identification. Machine Learning utilised to improve the performance of a multimodal biometric identification system. The performance evaluation is evaluated based on precision, recognition rate, equal error rate, and numerous evaluation metrics. The suggested multimodal system has an accuracy of 89.96 %, a false acceptance rate (FAR) of 3.32 %, and a false recognition rate (FRR) of 2.92 %. In order to arrive at this result, the multimodal system relies on score level fusion. It is demonstrated that a multimodal system may achieve high accuracy while using minimal FAR and FRR. [ABSTRACT FROM AUTHOR]

Real-time forecasting of the financial time-series data is challenging for many machine learning (ML) algorithms. First, many ML models operate offline, where they need a batch of data, which may not be available during training. Besides, due to a fixed architecture of the majority of the offline-based ML models, they suffer to deal with the uncertain nature of financial time-series data. In contrast, online learning mode evolving-structured ML models could be promising for financial time-series forecasting. For real-time deployment of such models, low memory demand is a must. Besides, the model’s explainability plays a crucial role in forecasting financial time-series. Considering all the requirements, a rule-based autonomous neuro-fuzzy learning algorithm called the parsimonious learning machine (PALM) is proposed here to forecast time-varying stock indices. To provide efficient automation of the proposed algorithm by maintaining the model explainability in terms of limited number linguistic IF-THEN rules, two popular multiobjective evolutionary algorithms (MEAs), such as a real-coded genetic algorithm (GA) and a self-adaptive differential evolution (DE) algorithm are utilized here. In addition, fuzzy type-2 variants of PALMs’ are considered here due to better uncertainty handling capacity than their type-1 counterparts. To evaluate the proposed algorithm’s performance, the closing stock price of fifteen (15) different stock market indices are predicted here. From the results, it is observed that the MEA-based PALMs are performing better than the state-of-the-art benchmark online ML models and providing a rule-based explainable model to the end-user. [ABSTRACT FROM AUTHOR]

Purpose: In this work, we propose an implementation of the Bienenstock–Cooper–Munro (BCM) model, obtained by a combination of the classical framework and modern deep learning methodologies. The BCM model remains one of the most promising approaches to modeling the synaptic plasticity of neurons, but its application has remained mainly confined to neuroscience simulations and few applications in data science. Methods: To improve the convergence efficiency of the BCM model, we combine the original plasticity rule with the optimization tools of modern deep learning. By numerical simulation on standard benchmark datasets, we prove the efficiency of the BCM model in learning, memorization capacity, and feature extraction. Results: In all the numerical simulations, the visualization of neuronal synaptic weights confirms the memorization of human-interpretable subsets of patterns. We numerically prove that the selectivity obtained by BCM neurons is indicative of an internal feature extraction procedure, useful for patterns clustering and classification. The introduction of competitiveness between neurons in the same BCM network allows the network to modulate the memorization capacity of the model and the consequent model selectivity. Conclusions: The proposed improvements make the BCM model a suitable alternative to standard machine learning techniques for both feature selection and classification tasks. [ABSTRACT FROM AUTHOR]

This paper deals with adaptive fuzzy approaches for controlling a cart-pendulum system. In fact, it suggests the use of a conventional Fuzzy Proportional-Integral-Derivative (FPID) controller to control the angle of the pendulum and an unconventional fuzzy controller which is an interesting idea to control the cart position. For that, two Adaptive FPID (AFPID) controllers and a Self-Tunable Fuzzy Inference System (STFIS) controller are used. To sweep all the output space of these controllers, a set-point very rich in information is applied to cart displacement. The conclusions of the fuzzy rules of STFIS controller are on-line fitted by using an adaptive learning algorithm. The conclusions obtained are sorted and then are clustered in intervals. Simulation results show the interpretability of STFIS controller behavior in form of decision rules. Moreover, the STFIS controller is compared to the two AFPID controllers. This comparison shows that a FIS based on a best self-tuning of fuzzy rule base, i.e. STFIS controller, can replace correctly and more perfectly the role of a FPID controller with self-tuning scaling factors. So, the STFIS controller might be used to improve performances and robustness of a control system. [ABSTRACT FROM AUTHOR]

Direct feedback alignment (DFA) is emerging as an efficient and biologically plausible alternative to backpropagation for training deep neural networks. Despite relying on random feedback weights for the backward pass, DFA successfully trains state-of-the-art models such as transformers. On the other hand, it notoriously fails to train convolutional networks. An understanding of the inner workings of DFA to explain these diverging results remains elusive. Here, we propose a theory of feedback alignment algorithms. We first show that learning in shallow networks proceeds in two steps: an alignment phase, where the model adapts its weights to align the approximate gradient with the true gradient of the loss function, is followed by a memorisation phase, where the model focuses on fitting the data. This two-step process has a degeneracy breaking effect: out of all the low-loss solutions in the landscape, a network trained with DFA naturally converges to the solution which maximises gradient alignment. We also identify a key quantity underlying alignment in deep linear networks: the conditioning of the alignment matrices. The latter enables a detailed understanding of the impact of data structure on alignment, and suggests a simple explanation for the well-known failure of DFA to train convolutional neural networks. Numerical experiments on MNIST and CIFAR10 clearly demonstrate degeneracy breaking in deep non-linear networks and show that the align-then-memorize process occurs sequentially from the bottom layers of the network to the top. [ABSTRACT FROM AUTHOR]

Asthma is a chronic lung disorder of which the number of sufferers estimated to be 10% of the population in Iran. Results of various studies show that asthma is usually under-diagnosed, especially in developing countries, because of limited access to medical specialist and laboratory data. The purpose of this paper is to design a Type-II fuzzy rule-based expert system to alleviate this hazard by diagnosing asthma at initial stages. Rule-base of this system is acquired from aggregation of patients' data and knowledge of experts. Results show that combination of these two methods can enhance the diagnosis capability of expert system. Type-II fuzzy technique is used to handle the uncertainty in this system and improve the accuracy of diagnosis. [ABSTRACT FROM AUTHOR]

In this study a new machine learning technique is presented to solve singular multi-pantograph differential equations (SMDEs). A new optimized type-3 fuzzy logic system (T3-FLS) by unscented Kalman filter (UKF) is proposed for solution estimation. The convergence and stability of presented algorithm are ensured by the suggested Lyapunov analysis. By two SMDEs the effectiveness and applicability of the suggested method is demonstrated. The statistical analysis show that the suggested method results in accurate and robust performance and the estimated solution is well converged to the exact solution. The proposed algorithm is simple and can be applied on various SMDEs with variable coefficients.

Chaos dynamics is an interesting nonlinear effect that can be observed in many chemical, electrical, and mechanical systems. The chaos phenomenon has many applications in various branches of engineering. On the other hand, the control of mobile robots to track unpredictable chaotic trajectories has a valuable application in many security problems and military missions. The main objective in this problem is to design a controller such that the robot tracks a desired chaotic path. In this paper, the concept of synchronization of chaotic systems is studied, and a new type-3 fuzzy system (T3FLS)-based controller is designed. The T3FLS is learned by some new adaptive rules. The new learning scheme of T3FLS helps to better stabilize and synchronize. The suggested controller has a better ability to cope with high-level uncertainties. Because, in addition to the fact that the T3FLSs have better ability in an uncertain environment, the designed compensator also improves the accuracy and robustness. Several simulations show better synchronization and control accuracy of the designed controller.

Machine learning is a method of data analysis, which allows the analytical system to learn in the course of solving many similar problems. Machine learning is based on the idea that analytical systems can learn how to identify patterns and make decisions with minimal human involvement. The history of already completed dialogues between users is used to train chat bots for automated communication with interlocutors. There are many machine learning algorithms, and this article describes the most popular of them and their use for teaching chat bots. [ABSTRACT FROM AUTHOR]

The machine learning approach has gained a significant attention in the healthcare sector because of the prospect of developing new techniques for medical devices and handling the critical database of chronic diseases. The learning approach has potential to analyze complex medical data, disease diagnosis, and patient monitoring system, and to monitor e-health record. Non-invasive cuffless blood pressure (CLBP) measurement secured a significant position in the patient monitoring system. From a few recent decades, the importance of cuffless technology has been perceived towards continuous monitoring of blood pressure (BP) and supplementary efforts have been made towards its continuous monitoring. However, the optimal method that measures BP unambiguously and continuously has not yet emerged along with issues like calibration time, accuracy and long-term estimation of BP with miniaturizing hardware. The present study provides an insight into several learning algorithms along with their feature selection models. Various challenges and future improvements towards the current state of machine learning in healthcare industries are discussed in the present review. The bottom line of this study is to provide a comprehensive perspective of the machine learning approach of CLBP for the generation of highly precise predictive models for continuous BP measurement. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel family of online censoring (OC) based complex-valued least mean kurtosis (CLMK) algorithms by inspiring the advantages of the online censoring strategy and kurtosis-based cost function. We first develop important members of this family of algorithms such as OC based CLMK (OC-CLMK) and augmented CLMK (OC-ACLMK). These algorithms censor less informative complex-valued data streams in an online manner and keep only the most informative ones for performing their weight vector updates. Thus, they reduce the costs of data processing without markedly affecting performance accuracy. However, they do not take into account possible outliers that disturb their performances considerably. Therefore, we also develop robust members such as robust OC-CLMK (ROC-CLMK) and OC-ACLMK (ROC-CLMK) algorithms, which censor less informative complex-valued data streams as well as possible outliers. The convergence analyses of the proposed algorithms in the sense of Lyapunov are also presented to derive the bounds of their step sizes in a compact form. The simulation results on large-scale system identification and regression scenarios affirm the mentioned attractive features of the proposed algorithms. This study also shows that the noteworthy properties of the proposed algorithms will provide important contributions to the processing of complex-valued big data streams. • A novel family of online censoring (OC) based CLMK algorithms is proposed. • The proposed OC-CLMK and OC-ACLMK censor less informative complex-valued data streams in an online manner. • They reduce data processing costs without markedly affecting performance accuracy. • We also develop robust members such as robust OC-CLMK and OC-ACLMK algorithms. • They censor both less informative complex-valued data streams and possible outliers. [ABSTRACT FROM AUTHOR]

In this study, a precise and efficient eigenvalue-based machine learning algorithm, particularly denoted as Eigenvalue Classification (EigenClass) algorithm, has been presented to deal with classification problems. The EigenClass algorithm is constructed by exploiting an eigenvalue-based proximity evaluation. To appreciate the classification performance of EigenClass, it is compared with the well-known algorithms, such as k-nearest neighbours, fuzzy k-nearest neighbours, random forest (RF) and multi-support vector machines. Number of 20 different datasets with various attributes and classes are used for the comparison. Every algorithm is trained and tested for 30 runs through 5-fold cross-validation. The results are then compared among each other in terms of the most used measures, such as accuracy, precision, recall, micro-F-measure, and macro-F-measure. It is demonstrated that EigenClass exhibits the best classification performance for 15 datasets in terms of every metric and, in a pairwise comparison, outperforms the other algorithms for at least 16 datasets in consideration of each metric. Moreover, the algorithms are also compared through statistical analysis and computational complexity. Therefore, the achieved results show that EigenClass is a precise and stable algorithm as well as the most successful algorithm considering the overall classification performances. [ABSTRACT FROM AUTHOR]

These days, because of the coronavirus, all countries are introducing online university systems. Online universities have the advantage of allowing students to take classes anytime, anywhere, 24 h a day, but lectures are given in a non-face-to-face manner between instructors and students. Thus, while students are taking classes on a web-based basis, the problem arises that concentration on the lectures is significantly reduced. Therefore, in order to solve these problems, in this paper, we propose a level-wise learning algorithm based on the difficulty level of the test problem, and we present the simulation results. In order to improve this problem, in this paper, we propose an automatic music recommendation algorithm based on fuzzy reasoning that can improve the level of learning and lecture concentration, and we show our results on developing a web-based, smart e-learning software. As a result of computer simulation, it was proved that the learning test method, considering by level the difficulty of the test and the incorrect answer rate, was more effective than the existing test method, judged the student's grades fairly, and improved the risk of unfairly failing the test by 30%. [ABSTRACT FROM AUTHOR]

This paper considers an autonomous cloud-based multi-robot system designed to execute highly repetitive tasks in a dynamic environment such as a modern megastore. Cloud level is intended for performing the most demanding operations in order to unload the robots that are users of cloud services in this architecture. For path planning on global level D* Lite algorithm is applied, bearing in mind its high efficiency in dynamic environments. In order to introduce smart cost map for further improvement of path planning in complex and crowded environment, implementation of fuzzy inference system and learning algorithm is proposed. The results indicate the possibility of applying a similar concept in different real-world robotics applications, in order to reduce the total paths length, as well as to minimize the risk in path planning related to the human-robot interactions. [ABSTRACT FROM AUTHOR]

Purpose: In this work, we propose an implementation of the Bienenstock–Cooper–Munro (BCM) model, obtained by a combination of the classical framework and modern deep learning methodologies. The BCM model remains one of the most promising approaches to modeling the synaptic plasticity of neurons, but its application has remained mainly confined to neuroscience simulations and few applications in data science. Methods: To improve the convergence efficiency of the BCM model, we combine the original plasticity rule with the optimization tools of modern deep learning. By numerical simulation on standard benchmark datasets, we prove the efficiency of the BCM model in learning, memorization capacity, and feature extraction. Results: In all the numerical simulations, the visualization of neuronal synaptic weights confirms the memorization of human-interpretable subsets of patterns. We numerically prove that the selectivity obtained by BCM neurons is indicative of an internal feature extraction procedure, useful for patterns clustering and classification. The introduction of competitiveness between neurons in the same BCM network allows the network to modulate the memorization capacity of the model and the consequent model selectivity. Conclusions: The proposed improvements make the BCM model a suitable alternative to standard machine learning techniques for both feature selection and classification tasks.

In this paper, a class of non-autonomous neural networks with time-varying delays is considered. By using a new differential inequality and M -matrix, we investigate the positive invariant set and global attracting set of the networks without the assumption on boundedness of time delays or system coefficients. On this basis, we obtain sufficient conditions on the uniformly boundedness, the existence of periodic attractor and give its existence range for periodic neural networks. Furthermore, we offer a weight learning algorithms to ensure input-to-state stability, and give the state estimate and attracting set for the system. Our results can extend and improve earlier ones. Some examples and simulations are given to demonstrate the effectiveness of the obtained results. [ABSTRACT FROM AUTHOR]

The paper presents a new approach to optimize the Multilayer Perceptron Neural Network (MLPNN) , to deal with the generalization problem. As known, most supervised learning algorithms aim to minimize the training error. However, the mentioned methods, based only on error minimizing, may generate a solution with an insufficient generalization performance. This present work proposes a multiobjective modelling problem involving two objectives: accuracy and complexity since the learning problem is multiobjective by nature. The learning task is carried on by minimizing both objectives simultaneously, according to Pareto domination concept, using NSGAII (Non-dominated Sorting Genetic Algorithm II) as a solver. This method leads us to a set of solutions called Pareto front, being the optimal solutions set, the adequate MLPNN need to be extracted. We show empirically that the proposed method is capable of reducing the neural networks topology and improved generalization performance, in addition to a good classification rate compared to different methods. [ABSTRACT FROM AUTHOR]

Three-phase transformers (TPT) play a significant and crucial function in the power networks in order to connect the sub-systems and deliver the electrical energy to final customers. The TPT are one of the most high-priced equipment in modern power networks, and therefore their working condition should be constantly monitored to prevent their breakdown, power outages and huge financial damage. Accordingly, this paper presents a hybrid method for detection and classification of incipient faults in TPT using dissolved gas analysis techniques (DGAT) information and rule-based machine learning method. In the developed method, the most informative and important items of DGAT data out of 14 items selected by association rules mining technique (ARMT) are employed as the input of adaptive neuro-fuzzy inference system (ANFIS). The ARMT is implemented to select the items, which have maximum information and can train the ANFIS more accurately. Furthermore, in order to enhance the accuracy of ANFIS and improve its robustness in different implementations, black widow optimization algorithm is applied for ANFIS training. In order to evaluate the performance of developed method on real issues, two industrial data collections obtained from Iran-Transfo Company chemical laboratory and Damavand power substations are used. The obtained results through MATLAB simulations proved that the developed method has high fault detection and classification accuracy, robust function in different implementations, short run time and simple structure. [ABSTRACT FROM AUTHOR]

Seven artificial neural network (ANN) models were developed to predict daytime actual evapotranspiration (ET) for Nissouri Creek in Oxford County, Canada, from April to July 2018, using the Bowen ratio energy balance method as target output for the first time. In total, 12 variations of each model were deployed using different combinations of model parameters, including the sigmoid and rectified linear unit (ReLU) activation functions, stochastic gradient descent (SGD), and root-mean-square-propagation (RMSprop) learning algorithms, three different network architectures, and 100 and 500 training epochs. This is the first time that ReLU has been used in ANNs that predict ET and it outperformed sigmoid in six of the seven models. This is particularly significant because until now the sigmoid activation function (or variations therein) had been exclusively employed in the ET literature. RMSprop was also used for the first time and typically demonstrated equivalent performance to that of SGD. The optimal model employs the ReLU activation function, consists of a 4-4-1 network architecture, includes the input parameters of net radiation, air temperature, soil heat flux, and wind speed, and is trained by the SGD learning algorithm for 500 training epochs. This model boasts a coefficient of determination (R2) of 0.997, root-mean-square error (RMSE) of 0.39 mm/day, and mean absolute error (MAE) of 0.18 mm/day. Furthermore, all seven models developed adequately model the ET process, with R2 ranging from 0.988 to 0.997, RMSE from 0.39 to 0.78 mm/day, and MAE from 0.18 to 0.58 mm/day. [ABSTRACT FROM AUTHOR]

Speech emotion recognition has become the heart of most human computer interaction applications in the modern world. The growing need to develop emotionally intelligent devices has opened up a lot of research opportunities. Most researchers in this field have applied the use of handcrafted features and machine learning techniques in recognising speech emotion. However, these techniques require extra processing steps and handcrafted features are usually not robust. They are computationally intensive because the curse of dimensionality results in low discriminating power. Research has shown that deep learning algorithms are effective for extracting robust and salient features in dataset. In this study, we have developed a custom 2D-convolution neural network that performs both feature extraction and classification of vocal utterances. The neural network has been evaluated against deep multilayer perceptron neural network and deep radial basis function neural network using the Berlin database of emotional speech, Ryerson audio-visual emotional speech database and Surrey audio-visual expressed emotion corpus. The described deep learning algorithm achieves the highest precision, recall and F1-scores when compared to other existing algorithms. It is observed that there may be need to develop customized solutions for different language settings depending on the area of applications. [ABSTRACT FROM AUTHOR]

Neural networks (NNs) are effective for the learning of nonlinear systems, and thus they achieve satisfactory results in various fields. However, they require significant amount of training data and learning time. Notably, the cerebellar model articulation controller (CMAC), which is modeled after the cerebellar neural transmission system, proposed by Albus can effectively reduce learning time, compared with NNs. The CMAC model is often used to learn nonlinear systems that have continuously changing outputs, i.e., regression problems. However, the structure of the CMAC model must be expanded to apply it to classification problems as well. Additionally, the CMAC model finds it difficult to simultaneously classify categories and estimate their proportional linear measure because designated learning algorithms are required for both regression and classification problems. Therefore, we aim to build a composite-type CMAC model that combines classification and regression algorithms to simultaneously classify categories and estimate their proportional linear measures. [ABSTRACT FROM AUTHOR]

In the past ten years, researchers have always attached great importance to the application of ontology to its relevant specific fields. At the same time, applying learning algorithms to many ontology algorithms is also a hot topic. For example, ontology learning technology and knowledge are used in the field of semantic retrieval and machine translation. The field of discovery and information systems can also be integrated with ontology learning techniques. Among several ontology learning tricks, multi-dividing ontology learning is the most popular one which proved to be in high efficiency for the similarity calculation of tree structure ontology. In this work, we study the multi-dividing ontology learning algorithm from the mathematical point of view, and an approximation conclusion is presented under the linear representation assumption. The theoretical result obtained here has constructive meaning for the similarity calculation and concrete engineering applications of tree-shaped ontologies. Finally, linear combination multi-dividing ontology learning is applied to university ontologies and mathematical ontologies, and the experimental results imply that the higher efficiency of the proposed approach in actual applications. [ABSTRACT FROM AUTHOR]

A conventional continuous passive motion (CPM) machine uses simple controller such as On/Off controller. Some better CPMs use PID controller. These kind of CPMs can not distinguish load different due to the different size of the patient leg. This may cause the CPM no longer follow the trajectory or the angle commands. Meanwhile, each patient may have different scenario of therapy from the others. When progress on the patient exists, the range of the flexion may be increased step by step. Therefore, the treatment can be different in term of the range of flexion from time to time. This paper proposes CPM with hybrid proportional integral derivative (PID) and iterative learning controller (ILC). The system has capability in learning the trajectory tracking. Therefore, the CPM will be able to follow any load or trajectory changes applied to it. The more accurate CPM machine can follow the trajectory command, the better its performance for the treatment. The experiment showed that the system was stable due to the PID controller. The tracking performance also improved with the ILC even there exist some disturbances. [ABSTRACT FROM AUTHOR]

Background: Optimizing programmed basal rates and carbohydrate ratios may improve the performance of the artificial pancreas. We tested, in a diabetes camp, the efficacy of a learning algorithm that updates daily basal rates and carbohydrate ratios in the artificial pancreas. Materials and Methods: We conducted a randomized crossover trial in campers and counselors aged 8-21 years with type 1 diabetes on pump therapy. Participants underwent 2 days of artificial pancreas alone and 6 days of artificial pancreas with learning. During the artificial pancreas with learning, programmed basal rates and carbohydrate ratios were updated daily based on the learning algorithm's recommendations. All algorithm recommendations were reviewed for safety by camp physicians. The primary outcome was the time in target range (3.9-10 mmol/L) of the last 2 days of each intervention. Results: Thirty-four campers (age 13.9 ± 3.9, hemoglobin A1c 8.3% ± 0.2%) were included. Ninety-six percent of algorithm recommendations were approved by the camp physicians. Participants were in closed-loop mode 74% of the time. There was no difference between interventions in time in target (55%-55%; P = 0.71) nor in hypoglycemia events (0.8-0.9 events per day; P = 0.63). This was despite changes in programmed basal rate ranging from -21% to +117%, and changes in breakfast, lunch, and supper carbohydrate ratios from -17% to +40%, -36% to +37%, and -35% to +63%, respectively. Morever, postprandial hyperglycemia and hypoglycemia did not decrease in participants whose carbohydrate ratios were decreased (more insulin boluses) and increased (less insulin boluses), respectively. Conclusions: In camp settings, despite adjustments to programmed basal rates and carbohydrate ratios, the learning algorithm did not change glycemia, which may point toward limited effect of these adjustments in environments with large day-to-day variability in insulin needs. Longer randomized studies in real-world settings are required to further assess the efficacy of automatic adjustments of programmed basal rates and carbohydrate ratios. [ABSTRACT FROM AUTHOR]

We introduce Hopfield-Energy-Based Learning, a general learning framework that is inspired by energy-based models, to train feedforward neural nets. Our approach includes two training phases applied iteratively: first, the minimization of the internal energy, which captures dependencies between input samples and network parameters, is carried out in an unsupervised manner; second, the problem-dependent supervised external energy (e.g., cross-entropy loss) combined with partially reversed internal energy gradients are back-propagated in a standard manner. The intuition is that the first stage helps parameters to settle into the state that simply partitions data into clusters; while in the second stage, the network is allowed to deviate from that clustering a bit (hence gradient reversal) in order to converge to parameters that ultimately perform well on the task at hand. Notably, the data used for the two steps might not be one and the same (e.g., can come from different domains) and the approach naturally tailors itself to solve unsupervised domain adaptation problems without adopting any distribution alignment techniques. We also show that the proposed training strategy substantially improves the performance of several ConvNets on standard supervised classification tasks; showing improvements of at least 1.2% (2.64% on CIFAR-10, 4.5% on CIFAR-100, and 1.35% on ImageNet). Our formulation is general, performs well in practice, and holds promise for scenarios where labeled data is limited. • We leverage Hopfield-based energy based learning to facilitate feedforward nets. • We introduce a mechanism to adapt compatibility of network parameters w.r.t inputs. • We use gradient reversal layer to control model energy without anti-Hebbian rule. • Our learning strategy outperforms UDA frameworks without any distribution alignment. • Our work has a universal value for network initialization and generalization. [ABSTRACT FROM AUTHOR]