Search

A novel evolutionary extreme learning machine (ELM) based on improved quantum-behaved particle swarm optimization (IQPSO) for radar target classification is presented in this paper. Quantum-behaved particle swarm optimization (QPSO) has been used in ELM to solve the problem that ELM needs more hidden nodes than conventional tuning-based learning algorithms due to the random set of input weights and hidden biases. But the method for calculating the characteristic length of Delta potential well of QPSO may reduce the global search ability of the algorithm. To solve this issue, a new method to calculate the characteristic length of Delta potential well is proposed in this paper. Experimental results based on the benchmark functions validate the better performance of IQPSO against QPSO in most cases. The novel algorithm is also evaluated by using real-world datasets and radar data; the experimental results indicate that the proposed algorithm is more effective than BP, SVM, ELM, QPSO-ELM, and so on, in terms of real-time performance and accuracy. [ABSTRACT FROM AUTHOR]

Particle swarm optimization (PSO) has long been attracting wide attention from researchers in the community. How to deal with the weak exploration ability and premature convergence of PSO remains an open question. In this paper, we modify the memory structure of canonical PSO and introduce the multi-leader mechanism to alleviate these problems. The proposed PSO variant in this paper is termed as multi-leader PSO (MLPSO) within which the modified memory structure provided more valuable information for particles to escape from the local optimum and multi-leader mechanism enhances diversity of particles' search pattern. Under the multi-leader mechanism, particles choose their leaders based on the game theory instead of a random selection. Besides, the best leader refers to other leaders' information to improve its quality in every generation based on a self-learning process. To make a comprehensive analysis, we test MLPSO against the benchmark functions in CEC 2013 and further applied MLPSO to a practical case: the reconstruction of gene regulatory networks based on fuzzy cognitive maps. The experimental results confirm that MLPSO enhances the efficiency of the canonical PSO and performs well in the realistic optimization problem. [ABSTRACT FROM AUTHOR]

This paper introduces a comprehensive and exhaustive overview of the cuckoo search algorithm (CSA). CSA is a metaheuristic swarm-based approach established by Yang and Deb [10] to emulate the cuckoo breeding behavior. Owing to the successful application of CSA for a wide variety of optimization problems, since then, researchers have developed several new algorithms in this field. This article displays a comprehensive review of all conducting intensive research survey into the pros and cons, main architecture, and extended versions of this algorithm. It is worth mentioning that the materials of this survey paper are categorized in accordance with the structure of the CSA in which the materials are divided into the CSA versions and modification, publication years, the CSA applications areas, and the hybridization of CSA. The survey paper ends with solid conclusions about the current research on CSA and the possible future directions for the relevant audience and readers. The researchers and practitioners on CSA belong to a wide range of audiences from the domains of optimization, engineering, medical, data mining, clustering, etc., who will benefit from this study. [ABSTRACT FROM AUTHOR]

Recently, deep recurrent neural networks have achieved great success in various machine learning tasks, and have also been applied for sound event detection. The detection of temporally overlapping sound events in realistic environments is much more challenging than in monophonic detection problems. In this paper, we present an approach to improve the accuracy of polyphonic sound event detection in multichannel audio based on gated recurrent neural networks in combination with auditory spectral features. In the proposed method, human hearing perception-based spatial and spectral-domain noise-reduced harmonic features are extracted from multichannel audio and used as high-resolution spectral inputs to train gated recurrent neural networks. This provides a fast and stable convergence rate compared to long short-term memory recurrent neural networks. Our evaluation reveals that the proposed method outperforms the conventional approaches. [ABSTRACT FROM AUTHOR]

The representation of consistent mixed reality (XR) environments requires adequate real and virtual illumination composition in real-time. Estimating the lighting of a real scenario is still a challenge. Due to the ill-posed nature of the problem, classical inverse-rendering techniques tackle the problem for simple lighting setups. However, those assumptions do not satisfy the current state-of-art in computer graphics and XR applications. While many recent works solve the problem using machine learning techniques to estimate the environment light and scene's materials, most of them are limited to geometry or previous knowledge. This paper presents a CNN-based model to estimate complex lighting for mixed reality environments with no previous information about the scene. We model the environment illumination using a set of spherical harmonics (SH) environment lighting, capable of efficiently represent area lighting. We propose a new CNN architecture that inputs an RGB image and recognizes, in real-time, the environment lighting. Unlike previous CNN-based lighting estimation methods, we propose using a highly optimized deep neural network architecture, with a reduced number of parameters, that can learn high complex lighting scenarios from real-world high-dynamic-range (HDR) environment images. We show in the experiments that the CNN architecture can predict the environment lighting with an average mean squared error (MSE) of 7.85 × 10−4 when comparing SH lighting coefficients. We validate our model in a variety of mixed reality scenarios. Furthermore, we present qualitative results comparing relights of real-world scenes. [Display omitted] • Automatic end-to-end method to estimate the environment lighting in XR applications. • A CNN architecture that learns a latent-space of the environment lighting. • A methodology to generate egocentric mixed-reality-views from HDR panoramas. • Real-time lighting estimation that does not make assumptions about the XR scene. [ABSTRACT FROM AUTHOR]

Convex optimization is explored in this paper to maximize the security against power analysis attacks while minimizing the power and performance overheads for a cryptographic circuit employs a random dynamic voltage and frequency scaling (RDVFS) technique. To realize this optimization goal, firstly, non-convex functions are used for modeling all the critical parameters related to the RDVFS technique to create a non-convex objective function. Subsequently, the non-convex objective function is appropriately converted into a convex objective function in order to grab the global optimum value. By transforming the acquired convex objective function into the Lagrange dual function and using Karush–Kuhn–Tucker (KKT) conditions to solve this optimization problem, the best RDVFS strategy can be uncovered. As shown in the result, after employing convex optimization for a cryptographic circuit with a RDVFS technique, the corresponding figure-of-merit (FOM) is enhanced by 43.4%. Moreover, in comparison with the conventional machine learning technique, the proposed convex optimization technique improves the overall FOM related with RDVFS technique by 17.6% while reducing 85.3% computational time. • Random dynamic voltage and frequency scaling (RDVFS) is beneficial for improving the robustness of cryptographic circuits. • RDVFS technique undermines the performance and power saving of cryptographic circuits. • Convex optimization can be utilized for improving the security of RDVFS technique without increasing the overall overhead significantly. • Convex optimization methodology has a better efficiency for optimizing RDVFS technique as compared to the training methodology related with MLP neural networks. [ABSTRACT FROM AUTHOR]

Many large-scale applications amount to finding relevant results from an enormous output space of potential candidates. For example, finding the best matching product from a large catalog or suggesting related search phrases on a search engine. The size of the output space for these problems can range from millions to billions, and can even be infinite in some applications. Moreover, training data is often limited for the long-tail items in the output space. Fortunately, items in the output space are often correlated thereby presenting an opportunity to alleviate the data sparsity issue. In this paper, we propose the Prediction for Enormous and Correlated Output Spaces (PECOS) framework, a versatile and modular machine learning framework for solving prediction problems for very large output spaces, and apply it to the eXtreme Multilabel Ranking (XMR) problem: given an input instance, find and rank the most relevant items from an enormous but fixed and finite output space. We propose a three phase framework for PECOS: (i) in the first phase, PECOS organizes the output space using a semantic indexing scheme, (ii) in the second phase, PECOS uses the indexing to narrow down the output space by orders of magnitude using a machine learned matching scheme, and (iii) in the third phase, PECOS ranks the matched items using a final ranking scheme. The versatility and modularity of PECOS allows for easy plug-and-play of various choices for the indexing, matching, and ranking phases. We also develop very fast inference procedures which allow us to perform XMR predictions in real time; for example, inference takes less than 1 millisecond per input on the dataset with 2.8 million labels. [ABSTRACT FROM AUTHOR]

Many important machine learning problems can be formulated as highly constrained convex optimization problems. One important example is metric constrained problems. In this paper, we show that standard optimization techniques can not be used to solve metric constrained problem. To solve such problems, we provide a general active set framework, called Project and Forget, and several variants thereof that use Bregman projections. Project and Forget is a general purpose method that can be used to solve highly constrained convex problems with many (possibly exponentially) constraints. We provide a theoretical analysis of Project and Forget and prove that our algorithms converge to the global optimal solution and have a linear rate of convergence. We demonstrate that using our method, we can solve large problem instances of general weighted correlation clustering, metric nearness, information theoretic metric learning and quadratically regularized optimal transport; in each case, out-performing the state of the art methods with respect to CPU times and problem sizes. [ABSTRACT FROM AUTHOR]

Contemporary learning models for computer vision are typically trained on very large (benchmark) datasets with millions of samples. These may, however, contain biases, artifacts, or errors that have gone unnoticed and are exploitable by the model. In the worst case, the trained model does not learn a valid and generalizable strategy to solve the problem it was trained for, and becomes a "Clever Hans" predictor that bases its decisions on spurious correlations in the training data, potentially yielding an unrepresentative or unfair, and possibly even hazardous predictor. In this paper, we contribute by providing a comprehensive analysis framework based on a scalable statistical analysis of attributions from explanation methods for large data corpora. Based on a recent technique — Spectral Relevance Analysis — we propose the following technical contributions and resulting findings: (a) a scalable quantification of artifactual and poisoned classes where the machine learning models under study exhibit Clever Hans behavior, (b) several approaches we collectively denote as Class Artifact Compensation, which are able to effectively and significantly reduce a model's Clever Hans behavior, i.e., we are able to un-Hans models trained on (poisoned) datasets, such as the popular ImageNet data corpus. We demonstrate that Class Artifact Compensation, defined in a simple theoretical framework, may be implemented as part of a neural network's training or fine-tuning process, or in a post-hoc manner by injecting additional layers, preventing any further propagation of undesired Clever Hans features, into the network architecture. Using our proposed methods, we provide qualitative and quantitative analyses of the biases and artifacts in, e.g., the ImageNet dataset, the Adience benchmark dataset of unfiltered faces, and the ISIC 2019 skin lesion analysis dataset. We demonstrate that these insights can give rise to improved, more representative, and fairer models operating on implicitly cleaned data corpora. [Display omitted] • We extend Spectral Relevance Analysis for large scale analyses of Clever Hans traits. • We introduce Class Artifact Compensation to unlearn or suppress confounders. • We show that Deep Neural Networks tend to overfit to Clever Hans artifacts in data. • Using both methods, we discover and unlearn Clever Hans artifacts on modern datasets. • We demonstrate that Class Artifact Compensation leads to more trustworthy models. [ABSTRACT FROM AUTHOR]

The graduation projects (GP) are important because it reflects the academic profile and achievement of the students. For many years' graduation projects are done by the information technology department students. Most of these projects have great value, and some were published in scientific journals and international conferences. However, these projects are stored in an archive room haphazardly and there is a very small part of it is a set of electronic PDF files stored on hard disk, which wastes time and effort and cannot benefit from it. However, there is no system to classify and store these projects in a good way that can benefit from them. In this paper, we reviewed some of the best machine learning algorithms to classify text "graduation projects", support vector machine (SVM) algorithm, logistic regression (LR) algorithm, random forest (RF) algorithm, which can deal with an extremely small amount of dataset after comparing these algorithms based on accuracy. We choose the SVM algorithm to classify the projects. Besides, we will mention how to deal with a super small dataset and solve this problem. [ABSTRACT FROM AUTHOR]

Support vector machines (SVMs) are among the most robust and accurate methods in all well-known machine learning algorithms, especially for classification. The SVMs train a classification model by solving an optimization problem to decide which instances in the training datasets are the support vectors (SVs). However, SVs are intact instances taken from the training datasets and directly releasing the classification model of the SVMs will carry significant risk to the privacy of individuals, when the training datasets contain sensitive information. In this paper, we study the problem of how to release the classification model of kernel SVMs while preventing privacy leakage of the SVs and satisfying the requirement of privacy protection. We propose a new differentially private algorithm for the kernel SVMs based on the exponential and Laplace hybrid mechanism named DPKSVMEL. The DPKSVMEL algorithm has two major advantages compared with existing private SVM algorithms. One is that it protects the privacy of the SVs by postprocessing and the training process of the non-private kernel SVMs does not change. Another is that the scoring function values are directly derived from the symmetric kernel matrix generated during the training process and does not require additional storage space and complex sensitivity analysis. In the DPKSVMEL algorithm, we define a similarity parameter to denote the correlation or distance between the non-SVs and every SV. And then, every non-SV is divided into a group with one of the SVs according to the maximal value of the similarity. Under some certain similarity parameter value, we replace every SV with a mean value of the top-k randomly selected most similar non-SVs within the group by the exponential mechanism if the number of non-SVs is greater than k. Otherwise, we add random noise to the SVs by the Laplace mechanism. We theoretically prove that the DPKSVMEL algorithm satisfies differential privacy. The extensive experiments show the effectiveness of the DPKSVMEL algorithm for kernel SVMs on real datasets; meanwhile, it achieves higher classification accuracy than existing private SVM algorithms. [ABSTRACT FROM AUTHOR]

At present, the graph neural network has achieved good results in the semisupervised classification of graph structure data. However, the classification effect is greatly limited in those data without graph structure, incomplete graph structure, or noise. It has no high prediction accuracy and cannot solve the problem of the missing graph structure. Therefore, in this paper, we propose a high-order graph learning attention neural network (HGLAT) for semisupervised classification. First, a graph learning module based on the improved variational graph autoencoder is proposed, which can learn and optimize graph structures for data sets without topological graph structure and data sets with missing topological structure and perform regular constraints on the generated graph structure to make the optimized graph structure more reasonable. Then, in view of the shortcomings of graph attention neural network (GAT) that cannot make full use of the graph high-order topology structure for node classification and graph structure learning, we propose a graph classification module that extends the attention mechanism to high-order neighbors, in which attention decays according to the increase of neighbor order. HGLAT performs joint optimization on the two modules of graph learning and graph classification and performs semisupervised node classification while optimizing the graph structure, which improves the classification performance. On 5 real data sets, by comparing 8 classification methods, the experiment shows that HGLAT has achieved good classification results on both a data set with graph structure and a data set without graph structure. [ABSTRACT FROM AUTHOR]

The simulated annealing algorithm is a stochastic optimization method for solving optimization problems associated with a large search space. The efficiency of the algorithm depends on the adaptation of the cooling model. Despite this, the main disadvantage of this algorithm is that it does not take into account the state of the system while searching. Thus, it is difficult to predict the system convergence with the simulated annealing algorithm. However, neural networks in particular continuous Hopfield networks have proven their ability in the field of machine learning to make a decision. In this paper, we introduce continuous Hopfield networks to improve the convergence of the simulated annealing algorithm. The experimental results show that the hybrid approach produces a large number of stable sets. [ABSTRACT FROM AUTHOR]

We propose the regularized recurrent inference machine (rRIM), a novel machine-learning approach to solve the challenging problem of deriving the pairing glue function from measured optical spectra. The rRIM incorporates physical principles into both training and inference and affords noise robustness, flexibility with out-of-distribution data, and reduced data requirements. It effectively obtains reliable pairing glue functions from experimental optical spectra and yields promising solutions for similar inverse problems of the Fredholm integral equation of the first kind. [ABSTRACT FROM AUTHOR]

Detecting anomalies in large networks is a major challenge. Nowadays, many studies rely on machine learning techniques to solve this problem. However, much of this research depends on synthetic or limited datasets and tends to use specialized machine learning methods to achieve good detection results. This study focuses on analyzing firewall logs from a large industrial control network and presents a novel method for generating anomalies that simulate real attacker actions within the network without the need for a dedicated testbed or installed security controls. To demonstrate that the proposed method is feasible and that the constructed logs behave as one would expect real-world logs to behave, different supervised and unsupervised learning models were compared using different feature subsets, feature construction methods, scaling methods, and aggregation levels. The experimental results show that unsupervised learning methods have difficulty in detecting the injected anomalies, suggesting that they can be seamlessly integrated into existing firewall logs. Conversely, the use of supervised learning methods showed significantly better performance compared to unsupervised approaches and a better suitability for use in real systems. [ABSTRACT FROM AUTHOR]

Field manual labor behavior recognition is an important task that applies deep learning algorithms to industrial equipment for capturing and analyzing people's behavior during field labor. In this study, we propose a field manual labor behavior recognition network based on an enhanced SlowFast architecture. The main work includes the following aspects: first, we constructed a field manual labor behavior dataset containing 433,500 fast-track frames and 8670 key frames based on the captured video data, and labeled it in detail; this includes 9832 labeled frames. This dataset provides a solid foundation for subsequent studies. Second, we improved the slow branch of the SlowFast network by introducing the combined CA (Channel Attention) attention module. Third, we enhanced the fast branch of the SlowFast network by introducing the ACTION hybrid attention module. The experimental results show that the recognition accuracy of the improved SlowFast network model with the integration of the two attention modules increases by 7.08%. This implies that the improved network model can more accurately locate and identify manual labor behavior in the field, providing a more effective method for problem solving. [ABSTRACT FROM AUTHOR]

Rate of penetration (ROP) is a key factor in drilling optimization, cost reduction and drilling cycle shortening. Due to the systematicity, complexity and uncertainty of drilling operations, however, it has always been a problem to establish a highly accurate and interpretable ROP prediction model to guide and optimize drilling operations. To solve this problem in the Tarim Basin, this study proposes four categories of hybrid physics-machine learning (ML) methods for modeling. One of which is residual modeling, in which an ML model learns to predict errors or residuals, via a physical model; the second is integrated coupling, in which the output of the physical model is used as an input to the ML model; the third is simple average, in which predictions from both the physical model and the ML model are combined; and the last is bootstrap aggregating (bagging), which follows the idea of ensemble learning to combine different physical models' advantages. A total of 5655 real data points from the Halahatang oil field were used to test the performance of the various models. The results showed that the residual modeling model, with an R2 of 0.9936, had the best performance, followed by the simple average model and bagging with R2 values of 0.9394 and 0.5998, respectively. From the view of prediction accuracy, and model interpretability, the hybrid physics-ML model with residual modeling is the optimal method for ROP prediction. [ABSTRACT FROM AUTHOR]

Conducting online estimation studies of the SOH of lithium-ion batteries is indispensable for extending the cycle life of energy storage batteries. Data-driven methods are efficient, accurate, and do not depend on accurate battery models, which is an important direction for battery state estimation research. However, the relationships between variables in lithium-ion battery datasets are mostly nonlinear, and a single data-driven algorithm is susceptible to a weak generalization ability affected by the dataset itself. Meanwhile, most of the related studies on battery health estimation are offline estimation, and the inability for online estimation is also a problem to be solved. In this study, an integrated learning method based on a stacking algorithm is proposed. In this study, the end voltage and discharge temperature were selected as the characteristics based on the sample data of NASA batteries, and the B0005 battery was used as the training set. After training on the dataset and parameter optimization using a Bayesian algorithm, the trained model was used to predict the SOH of B0007 and B0018 models. After comparative analysis, it was found that the prediction results obtained based on the proposed model not only have high accuracy and a short running time, but also have a strong generalization ability, which has a great potential to achieve online estimation. [ABSTRACT FROM AUTHOR]

In this paper, we provide external image features and use the internal attention mechanism to solve the VQA problem given a dataset of textual questions and related images. Most previous models for VQA use a pair of images and questions as input. In addition, the model adopts a question-oriented attention mechanism to extract the features of the entire image and then perform feature fusion. However, the shortcoming of these models is that they cannot effectively eliminate the irrelevant features of the image. In addition, the problem-oriented attention mechanism lacks in the mining of image features, which will bring in redundant image features. In this paper, we propose a VQA model based on adversarial learning and bidirectional attention. We exploit external image features that are not related to the question to form an adversarial mechanism to boost the accuracy of the model. Target detection is performed on the image—that is, the image-oriented attention mechanism. The bidirectional attention mechanism is conducive to promoting model attention and eliminating interference. Experimental results are evaluated on benchmark datasets, and our model performs better than other models based on attention methods. In addition, the qualitative results show the attention maps on the images and leads to predicting correct answers. [ABSTRACT FROM AUTHOR]

This paper features the study of global optimization problems and numerical methods of their solution. Such problems are computationally expensive since the objective function can be multi-extremal, nondifferentiable, and, as a rule, given in the form of a "black box". This study used a deterministic algorithm for finding the global extremum. This algorithm is based neither on the concept of multistart, nor nature-inspired algorithms. The article provides computational rules of the one-dimensional algorithm and the nested optimization scheme which could be applied for solving multidimensional problems. Please note that the solution complexity of global optimization problems essentially depends on the presence of multiple local extrema. In this paper, we apply machine learning methods to identify regions of attraction of local minima. The use of local optimization algorithms in the selected regions can significantly accelerate the convergence of global search as it could reduce the number of search trials in the vicinity of local minima. The results of computational experiments carried out on several hundred global optimization problems of different dimensionalities presented in the paper confirm the effect of accelerated convergence (in terms of the number of search trials required to solve a problem with a given accuracy). [ABSTRACT FROM AUTHOR]

Expert systems, a form of artificial intelligence (AI), are typically designed to solve many real-world problems by reasoning through knowledge, which is primarily represented as IF–THEN rules, with the information acquired from humans or domain experts. However, to assume such rules for personalized decision-making in an intelligent, context-aware mobile application is a challenging issue. The reason is that different mobile users may behave differently in various day-to-day situations, i.e., not identical, and thus the rules for personalized services must be reflected according to their symmetrical or asymmetrical behavioral activities. Therefore, our key focus is to solve this issue through adding personalized decision-making intelligence to develop powerful mobile applications to assist the end-users. To achieve our goal, in this paper, we explore on "Mobile Expert System", where we take into account machine-learning rules as knowledge-base rather than traditional handcrafted static rules. Thus, the concept of a mobile expert system enables the computing and decision-making processes more actionable and intelligent than traditional ones in the domain of mobile analytics and applications. Our experiment section shows that the context-aware machine learning rules discovered from users' mobile phone data can contribute in building a mobile expert system to solve a particular problem, through making personalized decisions in various context-aware test cases. [ABSTRACT FROM AUTHOR]

To beat competition and obtain valuable information, decision-makers must conduct in-depth machine learning or data mining for data analytics. Traditionally, clustering and classification are two common methods used in machine mining. For clustering, data are divided into various groups according to the similarity or common features. On the other hand, classification refers to building a model by given training data, where the target class or label is predicted for the test data. In recent years, many researchers focus on the hybrid of clustering and classification. These techniques have admirable achievements, but there is still room to ameliorate performances, such as distributed process. Therefore, we propose clustering and classification based on distributed automatic feature engineering (AFE) for customer segmentation in this paper. In the proposed algorithm, AFE uses artificial bee colony (ABC) to select valuable features of input data, and then RFM provides the basic data analytics. In AFE, it first initializes the number of cluster k. Moreover, the clustering methods of k-means, Wald method, and fuzzy c-means (FCM) are processed to cluster the examples in variant groups. Finally, the classification method of an improved fuzzy decision tree classifies the target data and generates decision rules for explaining the detail situations. AFE also determines the value of the split number in the improved fuzzy decision tree to increase classification accuracy. The proposed clustering and classification based on automatic feature engineering is distributed, performed in Apache Spark platform. The topic of this paper is about solving the problem of clustering and classification for machine learning. From the results, the corresponding classification accuracy outperforms other approaches. Moreover, we also provide useful strategies and decision rules from data analytics for decision-makers. [ABSTRACT FROM AUTHOR]

The data of the power Internet of Things (IOT) system is transferred from the IaaS layer to the SaaS layer. The general data preprocessing method mainly solves the problem of big data anomalies and missing at the PaaS layer, but it still lacks the ability to judge the high error data that meets the timing characteristics, making it difficult to deal with heterogeneous power inconsistent issues. This paper shows this phenomenon and its physical mechanism, showing the difficulty of building a quantitative model forward. A data-driven method is needed to form a hybrid model to correct the data. The research object is the electricity meter data on both sides of a commercial building transformer, which comes from different power IOT systems. The low-voltage side was revised based on the high-voltage side. Compared with the correction method based on purely using neural networks, the combined method, Linear Regression (LS) + Differential Evolution (DE) + Extreme Learning Machine (ELM), further reduces the deviation from approximately 4% to 1%. • The heterogeneous electrical data from complex IOT system is inconsistent. • The positive mechanism of inconsistency has been revealed, but not easy to correct. • Reverse data correction used a combination of simple methods and the effect is better. [ABSTRACT FROM AUTHOR]

In this paper, we propose a master–slave methodology to address the problem of optimal integration (location and sizing) of Distributed Generators (DGs) in Direct Current (DC) networks. This proposed methodology employs a parallel version of the Population-Based Incremental Learning (PPBIL) optimization method in the master stage to solve the location problem and the Vortex Search Algorithm (VSA) in the slave stage to solve the sizing problem. In addition, it uses the reduction of power losses as the objective function, considering all the constraints associated with the technical conditions specific to DGs and DC networks. To validate its effectiveness and robustness, we use as comparison methods, different solution methodologies that have been reported in the specialized literature, as well as two test systems (the 21 and 69-bus test systems). All simulations were performed in MATLAB. According to the results, the proposed hybrid (PPBIL–VSA) methodology provides the best trade-off between quality of the solution and processing times and exhibits an adequate repeatability every time it is executed. [ABSTRACT FROM AUTHOR]

Machine learning is a subcategory of artificial intelligence, which aims to make computers capable of solving complex problems without being explicitly programmed. Availability of large datasets, development of effective algorithms, and access to the powerful computers have resulted in the unprecedented success of machine learning in recent years. This powerful tool has been employed in a plethora of science and engineering domains including mining and minerals industry. Considering the ever-increasing global demand for raw materials, complexities of the geological structure of ore deposits, and decreasing ore grade, high-quality and extensive mineralogical information is required. Comprehensive analyses of such invaluable information call for advanced and powerful techniques including machine learning. This paper presents a systematic review of the efforts that have been dedicated to the development of machine learning-based solutions for better utilizing mineralogical data in mining and mineral studies. To that end, we investigate the main reasons behind the superiority of machine learning in the relevant literature, machine learning algorithms that have been deployed, input data, concerned outputs, as well as the general trends in the subject area. [ABSTRACT FROM AUTHOR]

In solving classification problems in the field of machine learning and pattern recognition, the pre-processing of data is particularly important. The processing of high-dimensional feature datasets increases the time and space complexity of computer processing and reduces the accuracy of classification models. Hence, the proposal of a good feature selection method is essential. This paper presents a new algorithm for solving feature selection, retaining the selection and mutation operators from traditional genetic algorithms. On the one hand, the global search capability of the algorithm is ensured by changing the population size, on the other hand, finding the optimal mutation probability for solving the feature selection problem based on different population sizes. During the iteration of the algorithm, the population size does not change, no matter how many transformations are made, and is the same as the initialized population size; this spatial invariance is physically defined as symmetry. The proposed method is compared with other algorithms and validated on different datasets. The experimental results show good performance of the algorithm, in addition to which we apply the algorithm to a practical Android software classification problem and the results also show the superiority of the algorithm. [ABSTRACT FROM AUTHOR]

Machine learning algorithms perform differently in settings with varying levels of training set mislabeling noise. Therefore, the choice of the right algorithm for a particular learning problem is crucial. The contribution of this paper is towards two, dual problems: first, comparing algorithm behavior; and second, choosing learning algorithms for noisy settings. We present the 'sigmoid rule' framework, which can be used to choose the most appropriate learning algorithm depending on the properties of noise in a classification problem. The framework uses an existing model of the expected performance of learning algorithms as a sigmoid function of the signal-to-noise ratio in the training instances. We study the characteristics of the sigmoid function using five representative non-sequential classifiers, namely, Naïve Bayes, kNN, SVM, a decision tree classifier, and a rule-based classifier, and three widely used sequential classifiers based on hidden Markov models, conditional random fields and recursive neural networks. Based on the sigmoid parameters we define a set of intuitive criteria that are useful for comparing the behavior of learning algorithms in the presence of noise. Furthermore, we show that there is a connection between these parameters and the characteristics of the underlying dataset, showing that we can estimate an expected performance over a dataset regardless of the underlying algorithm. The framework is applicable to concept drift scenarios, including modeling user behavior over time, and mining of noisy time series of evolving nature. [ABSTRACT FROM AUTHOR]

In this paper, we examine the use of ensemble methods in a multirobot task allocation environment. The aim is to enable a robot that needs to estimate the required resources to complete a task, to utilize information coming from other robots of the same type. To our knowledge, it is the first attempt made, to use such methods, to combine data of the same type, coming from data sets of different agents, to form a prediction. Knowledge exchange is not continuous, but only ad hoc. To merge data, we use ensemble models. This keeps communication needs to a minimum, as only the models themselves-and no actual data- need to be exchanged. To further reduce communication costs, the number of robots that contribute information is being limited. Finally, we make an attempt to see how well the concept we use would perform in other domains. This is to examine whether the approach could yield the same results in other domains, or it is limited to the task allocation problem, as formulated in Tolmidis and Petrou ( Eng. Appl. Artif. Intell., 2013;26(5-6):1458-1468) . For this, we selected two additional, different, publicly available data sets. [ABSTRACT FROM AUTHOR]

High prediction accuracies are not the only objective to consider when solving problems using machine learning. Instead, particular scientific applications require some explanation of the learned prediction function. For computational biology, positional oligomer importance matrices (POIMs) have been successfully applied to explain the decision of support vector machines (SVMs) using weighted-degree (WD) kernels. To extract relevant biological motifs from POIMs, the motifPOIM method has been devised and showed promising results on real-world data. Our contribution in this paper is twofold: as an extension to POIMs, we propose gPOIM, a general measure of feature importance for arbitrary learning machines and feature sets (including, but not limited to, SVMs and CNNs) and devise a sampling strategy for efficient computation. As a second contribution, we derive a convex formulation of motifPOIMs that leads to more reliable motif extraction from gPOIMs. Empirical evaluations confirm the usefulness of our approach on artificially generated data as well as on real-world datasets. [ABSTRACT FROM AUTHOR]

Creating sessions in scientific conferences consists in grouping papers with common topics taking into account the size restrictions imposed by the conference schedule. Therefore, this problem can be considered as semi-supervised clustering of documents based on their content. This paper aims to propose modifications in traditional clustering algorithms to incorporate size constraints in each cluster. Specifically, two new algorithms are proposed to semi-supervised clustering, based on: binary integer linear programming with cannot-link constraints and a variation of the K-Medoids algorithm, respectively. The applicability of the proposed semi-supervised clustering methods is illustrated by addressing the problem of automatic configuration of conference schedules by clustering articles by similarity. We include experiments, applying the new techniques, over real conferences datasets: ICMLA-2014, AAAI-2013 and AAAI-2014. The results of these experiments show that the new methods are able to solve practical and real problems. [ABSTRACT FROM AUTHOR]

Most of the proposed algorithms to solve the dynamic clustering problem are based on nature inspired meta-heuristic algorithms. In this paper a different reinforcement based optimization approach called continuous action-set learning automata (CALA) is used and a novel dynamic clustering approach called ACCALA is proposed. CALA is an optimization tool interacting with a random environment and learn the optimal action from the environment feedbacks. In this paper the dynamic clustering problem considered as a noisy optimization problem and the team of CALAs is used to solve this noisy optimization problem. To build such a team of CALAs this paper proposed a new representation of CALAs. Each automaton in this team uses its continuous action-set and defining a suitable action-set for each automaton has a great impact on the CALAs search behavior. In this paper we used the statistical property of data-sets and proposed a new method to automatically find an action-set for each automaton. The performance of ACCALA is evaluated and the results are compared with seven well-known automatic clustering techniques. Also ACCALA is used to perform automatic segmentation. The experimental results are promising and show that the proposed algorithm produced compact and well-separated clusters. [ABSTRACT FROM AUTHOR]

Machine Reassignment is a challenging problem for constraint programming (CP) and mixed integer linear programming (MILP) approaches, especially given the size of data centres. Hybrid solutions mixing CP and heuristic algorithms, such as, large neighbourhood search (CBLNS), also struggle to address the problem given its size and number of constraints. The multi-objective version of the Machine Reassignment Problem is even more challenging and it seems unlikely for CP, MILP or hybrid solutions to obtain good results in this context. As a result, the first approaches to address this problem have been based on other optimisation methods, including metaheuristics. In this paper we study three things: (i) under which conditions a mixed integer optimisation solver, such as IBM ILOG CPLEX, can be used for the Multi-objective Machine Reassignment Problem; (ii) how much of the search space can a well-known hybrid method such as CBLNS explore; and (iii) can we find a better hybrid approach combining MILP or CBLNS and an- other recent metaheuristic proposed for the problem (GeNePi). We show that MILP can handle only small or medium scale data centres, and with some relaxations, such as, an optimality tolerance gap and a limited number of directions explored in the search space. CBLNS on the other hand struggles with the problem in general but achieves reasonable performance for large instances of the problem. However, we show that our hybridisation improves both the quality of the set of solutions (CPLEX+GeNePi and CBLNS+GeNePi improve the solutions by +17.8% against CPLEX alone and +615% against CBLNS alone) and number of solutions (8.9 times more solutions than CPLEX alone and 56.76 times more solutions than CBLNS alone), while the processing time of CPLEX+GeNePi and CBLNS+GeNePi increases only by 6% and 16.4% respectively. Overall, the study shows that CPLEX+GeNePi is the best algorithm for small instances (CBLNS+GeNePi only gets 45.2% of CPLEX+GeNePi's hypervolume) while CBLNS+GeNePi is better than the others on large instances (that CPLEX+GeNePi cannot address). [ABSTRACT FROM AUTHOR]

In this paper, a graph-based optimization problem related to epidemics control is studied. This multiobjective optimization problem requires determining which graph nodes to protect in order to limit the number of nodes affected by a threat spreading in the graph. In the paper, a classifier-based evolutionary algorithm is proposed in which population initialization, crossover, mutation and local search can use a previously trained classifier to decide which graph nodes to protect. The classifier is trained on a dataset collected by running an optimization algorithm on small problem instances and it is subsequently reused for solving multiple, larger instances of the same optimization problem. Selecting graph nodes to protect using a previously trained classifier allows the algorithm to find better solutions to the optimization problem compared to an algorithm not using a classifier. In the experiments, several machine learning models were evaluated and a multilayer perceptron (MLP) neural network was selected as the best performing classifier. The proposed optimization method was tested and found to work effectively when the MLP classifier was trained on data collected when solving problem instances with 1000 graph nodes. In tests on problem instances with up to 20000 nodes the proposed method outperformed an evolutionary optimizer not using any classifier-based components thereby showing that knowledge can be transferred from smaller problem instances to larger ones. [Display omitted] • A graph-based optimization problem related to epidemics control is studied. • A classifier-based evolutionary multiobjective optimization algorithm is proposed. • A multilayer perceptron (MLP) neural network is the best performing classifier. • The proposed method outperforms an evolutionary optimizer not using any classifier-based components. • Results show that knowledge can be transferred from smaller problem instances to larger ones. [ABSTRACT FROM AUTHOR]

Truecasing is the process of finding the proper capitalization of words in a text where the correct case information of the word is not given. The case of letters is often lost while performing language translation from non-capitalized languages (eg: Japanese, Arabic, Hindi) to capitalized languages (eg: Greek, Latin, English). In Optical Character Recognition (OCR) where the different cases for letters like 'p' and 'P' are similar, the case as well as the context and proper meaning of the text is lost. A similar situation is observed for text generated from Automatic Speech Recognition. This paper proposes to solve the problem of truecasing while focusing on three different architectures, all based on transformers. Transformers have generally proved to outperform recurrent and convolutional models. A proposal for a novel light-weight transformer architecture has been made which performs competitively with existing models that are computationally expensive. To provide a basis for comparison, two existing transformer architectures were fine-tuned for truecasing - BERT and GPT-2. These architectures are analyzed and compared on various metrics. The proposed architecture shows an accuracy of 98.25%, which is also an improvement of 0.41 against the current state-of-art. [ABSTRACT FROM AUTHOR]

The paper presents an approach to machine learning involving the use of a hybrid method in Neural ODE. The authors consider a way of the Neural ODE application with an embedded hybrid method for intelligent manipulator control in a non-deterministic environment. To reduce computational resources, the number of layers of the neural network using ODE is reduced, the solution of which is carried out by a hybrid transformation method. The hybrid method includes normalizing the system, linear transformation, and polynomial transformation. Using the hybrid method in Neural ODE can increase the learning and computation speed. The transformation method has high accuracy with low computational complexity of the algorithm. As a result of applying the hybrid method, the original nonlinear system is transformed to a simpler form. To solve the simplified transformed system, the scheme of the numerical Runge-Kutta method of the third order with an adaptive step is used. The results of the computational experiment show an increase in the speed and accuracy of the calculation. The presented approach to implementing the hybrid method in Neural ODE allows faster training of the neural network of the world model on limited data obtained from the real environment. After training on the data generated by the model of the world, a robotic arm always successfully reaches the object during subsequent interaction with the real environment. This solves the problem of the limited ability of the agent to interact when learning with the real environment in robotic systems. [ABSTRACT FROM AUTHOR]

Crime is a huge problem in our society, and fighting it is crucial. On a daily basis, there are several crimes done on a massive scale. For future reference, a database needs to be kept of all criminal acts and acts of misconduct, as these events are unlikely to be reported later. One of the issues now encountered is keeping up with crime statistics as well as analyzing this data to assist in preventing and solving future crimes. The aim of this project is to gather and analyze datasets containing a multitude of criminal events, and to determine which kinds of criminal activity may occur in the future based on various factors. We will use Machine learning algorithms for crime prediction in Chicago by applying machine learning algorithms to the Chicago criminal data set. Decision Tree, Gaussian Naive Bayes, k-NN, Logistic Regression are all excellent supervised classification tools for this task. With this method, we are attempting to predict crimes by classifying, recognizing patterns, and using effective technologies and tools. We can correlate aspects which can aid our comprehension of future crime patterns by looking at crime data trends that have occurred in the past. Machine learning and visualization approaches are applied to estimating that crimes dispersion over an area in this project. Once the primary records are analyzed and shown based on the need, the procedure shifted to the next stage. A three-dimensional modeling of Cologne's city is constructed using crime statistics. Result visualization in geovirtual environments and real-time situation monitoring in a social protection context result from exploring crimes data analysis in geovirtual environments. Three-dimensional measurements are rapidly becoming the norm in the documenting procedure for crime scenes. By utilizing 3D measuring tools, a richer investigation is provided, revealing the presence of each piece of evidence in relation to the rest of the crime scene. [ABSTRACT FROM AUTHOR]

Machine learning provides an effective approach to execute big data analysis. As a branch of machine learning, classification has been widely adopted in data processing. However, the sensitivity of data raises the concern of data privacy. How to balance data utility and data privacy is a challenging issue. Privacy-preserving data classification, which supports flexible and privacy-friendly access to datasets and data classification, enables users' data to be collected in an authenticated manner. However, the priva-cy-preserving data classification approach has a limitation in that the correctness of data classification cannot be guaranteed. As a consequence, it is possible for a malicious classifier to manipulate the classification result. To solve these problems, in this article, we propose a transparent and accountable privacy-preserving data classification framework, which involves a tracer to assert the behavior of the classifier and maintains the utility and privacy of data. Specifically, we take advantage of cryptography techniques to balance data privacy and data utility, and use blockchain to achieve transparency and accountability for the behavior of the classifier. To illustrate the practicability of this framework, we implement concrete cryptographic algorithms and develop a prototype system to evaluate and test its performance. [ABSTRACT FROM AUTHOR]

Complex systems are usually affected by various sources of uncertainty, and it is essential to account for mechanisms that ensure the proper management of such disturbances. This paper introduces a novel approach to solve symbolic regression problems, which combines the potential of Grammatical Evolution to obtain solutions by describing the search space with context-free grammars, and the ability of Modal Interval Analysis (MIA) to handle quantified uncertainty. The presented methodology uses an MIA solver to evaluate the fitness function, which represents a novel method to manage uncertainty by means of interval-based prediction models. This paper first introduces the theory that establishes the basis of the proposed methodology, and follows with a description of the system architecture and implementation details. Then, we present an illustrative application example which consists of determining the outer and inner approximations of the mean velocity of the water current of a river stretch. Finally, the interpretation of the obtained results and the limitations of the proposed methodology are discussed. [ABSTRACT FROM AUTHOR]

The KNN algorithm is one of the most famous algorithms in machine learning and data mining. It does not preprocess the data before classification, which leads to longer time and more errors. To solve the problems, this paper first proposes a PK-means++ algorithm, which can better ensure the stability of a random experiment. Then, based on it and spherical region division, an improved KNNPK+ is proposed. The algorithm can select the center of the spherical region appropriately and then construct an initial classifier for the training set to improve the accuracy and time of classification. [ABSTRACT FROM AUTHOR]

Intrusion detection system (IDS), the second security gate behind the firewall, can monitor the network without affecting the network performance and ensure the system security from the internal maximum. Many researches have applied traditional machine learning models, deep learning models, or hybrid models to IDS to improve detection effect. However, according to Predicted accuracy, Descriptive accuracy, and Relevancy (PDR) framework, most of detection models based on model-based interpretability lack good detection performance. To solve the problem, in this paper, we have proposed a novel intrusion detection system model based on model-based interpretability, called Interpretable Intrusion Detection System (I2DS). We firstly combine normal and attack samples reconstructed by AutoEncoder (AE) with training samples to highlight the normal and attack features, so that the classifier has a gorgeous effect. Then, Additive Tree (AddTree) is used as a binary classifier, which can provide excellent predictive performance in the combined dataset while maintaining good model-based interpretability. In the experiment, UNSW-NB15 dataset is used to evaluate our proposed model. For detection performance, I2DS achieves a detection accuracy of 99.95%, which is better than most of state-of-the-art intrusion detection methods. Moreover, I2DS maintains higher simulatability and captures the decision rules easily. [ABSTRACT FROM AUTHOR]

In SDN anomaly detection systems, when a training mechanism adopts semi-supervised learning (consisting of self-training and self-learning) to attain the classifiers of online training, it may cause the accumulation of identification errors - to degrade the performance. This paper presents a new training and learning mechanism which involves the operations of self-training and active learning to solve the problem. The proposed mechanism first adds samples with "high confidence weights" and classified as "malicious" to the training set by random selection. It then practices active learning to label those samples with "low confidence weights" and add them to the training set for training, to further lift up identification accuracy. A faster clustering method has been brought in to reduce the operation time of active learning. In classifier retraining, parallel training is applied to keep the classifier in constant service without interruption. Simulation results show that, in contrast to existing active learning IDS (ALIDS), our new mechanism performs better in identifying unknown attacks, without occupying the operation time of detection as it processes both training and detection in parallel. [ABSTRACT FROM AUTHOR]

Nearest neighbor-based methods are classic techniques that, due to their efficiency, still are widely used today. However, they have not been broadly applied to solve the multi-instance multi-label (MIML) problem, a supervised learning paradigm that combines multi-instance (MI) and multi-label (ML) learning. This work presents new neighbor-based approaches for solving MIML problems. On the one hand, MIML data are transformed into ML data and ML nearest neighbor algorithms are used. On the other hand, algorithms that directly address MIML data and use a bag-based distance are proposed. A comprehensive study and an overall comparison have been conducted to study the performance of these methods using different configurations. Experiments included 16 datasets and 8 performance metrics. The results and statistical tests showed that the problem transformation applied and the distance function used impacted the performance and that the approaches that do not transform the problem obtained the best predictive results. Furthermore, most of the proposed algorithms outperformed the MIMLkNN algorithm, the state-of-art algorithm for MIML learning based on nearest-neighbor. Therefore, the relevance and capabilities of neighbor-based approaches to obtain competitive results in MIML learning are shown. Finally, all the algorithms developed in this paper have been included in the MIML library to facilitate the comparison with other future proposals. • New kNN methods for MIML based on logistic regression, gravitational model, and MAP. • Exhaustive study with 16 datasets and 8 evaluation measures to study the performance. • The transformation and the distance function influenced the predictive performance. • The gravitational model gets the best results with reasonable time complexity. • All the proposals outperform the state-of-art kNN-based algorithm in MIML. [ABSTRACT FROM AUTHOR]

The design and functionality of the human–machine interface (HMI) significantly affects operational efficiency and safety related to process control. Alarm management techniques consider the cognitive model of operators, but mainly only from a signal perception point of view. To develop a human-centric alarm management system, the construction of an easy-to-use and supportive HMI is essential. This work suggests a development method that uses machine learning (ML) tools. The key idea is that more supportive higher-level HMI displays can be developed by analysing operator-related events in the process log file. The obtained process model contains relevant data on the relationship of the process events, enabling a network-like visualisation. Attributes of the network allow us to solve the minimisation problem of the ideal workflow–display relation. The suggested approach allows a targeted process pattern exploration to design higher-level HMI displays with respect to content and hierarchy. The method was applied in a real-life hydrofluoric acid alkylation plant, where a proposal was made about the content of an overview display. [ABSTRACT FROM AUTHOR]

For text big data analysis, we preprocessed text data and constructed a document–keyword matrix. The elements of this matrix represent the frequencies of keywords occurring in a document. The matrix has a zero-inflation problem because many elements are zero values. Also, in the process of preprocessing, the data size of the document–keyword matrix is reduced. However, various machine learning algorithms require a large amount of data, so to solve the problems of data shortage and zero inflation, we propose the use of generative models based on statistics and machine learning. In our experimental tests, we compared the performance of the models using simulation and practical data sets. Thus, we verified the validity and contribution of our research for keyword data analysis. [ABSTRACT FROM AUTHOR]

In a vast majority of cases, remediation of IT issues encoded into domain-specific or user-defined alerts occurring in cloud environments and customer ecosystems suffers from accurate recommendations, which could be supplied in a timely manner for recovery of performance degradations. This is hard to realize by furnishing those abnormality definitions with appropriate expert knowledge, which varies from one environment to another. At the same time, in many support cases, the reported problems under Global Support Services (GSS) or Site Reliability Engineering (SRE) treatment ultimately go down to the product teams, making them waste costly development hours on investigating self-monitoring metrics of our solutions. Therefore, the lack of a systematic approach to adopting AI Ops significantly impacts the mean-time-to-resolution (MTTR) rates of problems/alerts. This would imply building, maintaining, and continuously improving/annotating a data store of insights on which ML models are trained and generalized across the whole customer base and corporate cloud services. Our ongoing study aligns with this vision and validates an approach that learns the alert resolution patterns in such a global setting and explains them using interpretable AI methodologies. The knowledge store of causative rules is then applied to predicting potential sources of the application degradation reflected in an active alert instance. In this communication, we share our experiences with a prototype solution and up-to-date analysis demonstrating how root conditions are discovered accurately for a specific type of problem. It is validated against the historical data of resolutions performed by heavy manual development efforts. We also offer experts a Dempster–Shafer theory-based rule verification framework as a what-if analysis tool to test their hypotheses about the underlying environment. [ABSTRACT FROM AUTHOR]

Quantum computing has rapidly gained prominence for its unprecedented computational efficiency in solving specific problems when compared to classical computing counterparts. This surge in attention is particularly pronounced in the realm of quantum machine learning (QML) following a classical trend. Here we start with a comprehensive overview of the current state-of-the-art in Quantum Support Vector Machines (QSVMs). Subsequently, we analyze the limitations inherent in both annealing and gate-based techniques. To address these identified weaknesses, we propose a novel hybrid methodology that integrates aspects of both techniques, thereby mitigating several individual drawbacks while keeping the advantages. We provide a detailed presentation of the two components of our hybrid models, accompanied by the presentation of experimental results that corroborate the efficacy of the proposed architecture. These results pave the way for a more integrated paradigm in quantum machine learning and quantum computing at large, transcending traditional compartmentalization. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI) has emerged as a crucial tool in healthcare with the primary aim of improving patient outcomes and optimizing healthcare delivery. By harnessing machine learning algorithms, natural language processing, and computer vision, AI enables the analysis of complex medical data. The integration of AI into healthcare systems aims to support clinicians, personalize patient care, and enhance population health, all while addressing the challenges posed by rising costs and limited resources. As a subdivision of computer science, AI focuses on the development of advanced algorithms capable of performing complex tasks that were once reliant on human intelligence. The ultimate goal is to achieve human-level performance with improved efficiency and accuracy in problem-solving and task execution, thereby reducing the need for human intervention. Various industries, including engineering, media/entertainment, finance, and education, have already reaped significant benefits by incorporating AI systems into their operations. Notably, the healthcare sector has witnessed rapid growth in the utilization of AI technology. Nevertheless, there remains untapped potential for AI to truly revolutionize the industry. It is important to note that despite concerns about job displacement, AI in healthcare should not be viewed as a threat to human workers. Instead, AI systems are designed to augment and support healthcare professionals, freeing up their time to focus on more complex and critical tasks. By automating routine and repetitive tasks, AI can alleviate the burden on healthcare professionals, allowing them to dedicate more attention to patient care and meaningful interactions. However, legal and ethical challenges must be addressed when embracing AI technology in medicine, alongside comprehensive public education to ensure widespread acceptance. [ABSTRACT FROM AUTHOR]

Non-convex optimization, which can better capture the problem structure, has received considerable attention in the applications of machine learning, image/signal processing, statistics, etc. With faster convergence rate, there have been tremendous studies on developing stochastic variance reduced algorithms to solve these non-convex optimization problems. However, as a crucial hyper-parameter for stochastic variance reduced algorithms, that how to select an appropriate step size is less researched in solving non-convex optimization problems. To address this gap, we propose a new class of stochastic variance reduced algorithms based on hyper-gradient, which has the ability to automatically obtain the online step size. Specifically, we focus on the variance-reduced stochastic optimization algorithms, the stochastic variance reduced gradient (SVRG) algorithm, which computes a full gradient periodically. We analyze theoretically the convergence of the proposed algorithm for non-convex optimization problems. Moreover, we show that the proposed algorithm enjoys the same complexities as state-of-the-art algorithms for solving non-convex problems in terms of finding an approximate stationary point. Thorough numerical results on empirical risk minimization with non-convex loss functions validate the efficacy of our method. [ABSTRACT FROM AUTHOR]