Search

The confluence of significant computational power and inexpensive sensors provides the opportunity to reliably collect large volumes of information from the world and extract humanly useful information resources. This paper reviews a coherent body of work over the last 20+ years focused on development of advanced bio-inspired computing techniques, and their applications primarily for human related data in behaviour and human centered computing. We close with a synthesis proposing an experiment analysis methodology combining these tools.

With the speedy progress of network technology and internet applications, protection of digitized data against unauthorized access and modification has become a paramount issue. In order to face this challenge, numerous secret image sharing schemes have been introduced. Secret image sharing scheme is a method used for safeguarding the sensitive digitized image against illegal copying and tempering. The secret image is divided into many random shares in such a way that each share does not reveal any information to the intruders. In this paper, we present a comprehensive survey from over 100 papers which explains the new approaches and challenges. This paper also provides a comparative analysis of different methods based on different properties.

In the paper we present an approach to the automatic detection and identification of important elements in paper documents. This includes stamps, logos, printed text blocks, signatures and tables. Presented approach consists of two stages. The first one includes object detection by means of AdaBoost cascade of weak classifiers and Haar-like features. Resulting image blocks are, at the second stage, subjected to verification based on selected features calculated from recently proposed low-level descriptors combined with certain classifiers representing current machine-learning approaches. The training phase, for both stages, uses bootstrapping, i.e., integrative process, aiming at increasing the accuracy. Experiments performed on large set of digitized paper documents showed that adopted strategy is useful and efficient.

Exploring massive mobile data for location-based services becomes one of the key challenges in mobile data mining. In this paper, we investigate a problem of finding a correlation between the collective behavior of mobile users and the distribution of points of interest (POIs) in a city. Specifically, we use large-scale cell tower data dumps collected from cell towers and POIs extracted from a popular social network service, Weibo. Our objective is to make use of the data from these two different types of sources to build a model for predicting the POI densities of different regions in the covered area. An application domain that may benefit from our research is a business recommendation application, where a prediction result can be used as a recommendation for opening a new store/branch. The crux of our contribution is the method of representing the collective behavior of mobile users as a histogram of connection counts over a period of time in each region. This representation ultimately enables us to apply a supervised learning algorithm to our problem in order to train a POI prediction model using the POI data set as the ground truth. We studied 12 state-of-the-art classification and regression algorithms; experimental results demonstrate the feasibility and effectiveness of the proposed method.

Recently, virus diseases, such as SARS-CoV, MERS-CoV, and COVID-19, continue to emerge and pose a severe public health problem. These diseases threaten the lives of many people and cause serious social and economic losses. Recent developments in information technology (IT) and connectivity have led to the emergence of Internet of Things (IoT) and Artificial Intelligence (AI) applications in many industries. These industries, where IoT and AI together are making significant impacts, are the healthcare and the diagnosis department. In addition, by actively communicating with smart devices and various biometric sensors, it is expanding its application fields to telemedicine, healthcare, and disease prevention. Even though existing IoT and AI technologies can enhance disease detection, monitoring, and quarantine, their impact is very limited because they are not integrated or applied rapidly to the emergence of a sudden epidemic. Especially in the situation where infectious diseases are rapidly spreading, the conventional methods fail to prevent large-scale infections and block global spreads through prediction, resulting in great loss of lives. Therefore, in this paper, various sources of infection information with local limitations are collected through virus disease information collector, and AI analysis and severity matching are performed through AI broker. Finally, through the Integrated Disease Control Center, risk alerts are issued, proliferation block letters are sent, and post-response services are provided quickly. Suppose we further develop the proposed integrated virus disease control model. In that case, it will be possible to proactively detect and warn of risk factors in response to infectious diseases that are rapidly spreading worldwide and strengthen measures to prevent spreading of infection in no time.

A prediction model is an indispensable tool in business, helping to make decisions, whether in the short, medium, or long term. In this context, the implementation of machine learning techniques in time series forecasting models has a notorious relevance, as information processing and efficient and dynamic knowledge uncovering are increasingly demanded. This paper develops a model called Variable step-size evolving Participatory Learning with Kernel Recursive Least Squares, VS-ePL-KRLS, applied to the forecast of weekly prices for S500 and S10 diesel oil, at the Brazilian level, for biweekly and monthly horizons. The presented model demonstrates a better accuracy compared with analogous models in the literature, without loss of computational performance for all time series analyzed.

Diabetic retinopathy (DR) is a significant reason for the global increase in visual loss. Studies show that timely treatment can significantly bring down such incidents. Hence, it is essential to distinguish the stages and severity of DR to recommend needed medical attention. In this view, this paper presents DRISTI (Diabetic Retinopathy classIfication by analySing reTinal Images), where a hybrid deep learning model composed of VGG16 and capsule network is proposed, which yields statistically significant performance improvement over the state of the art. To validate our claim, we have reported detailed experimental and ablation studies. We have also created an augmented dataset to increase the APTOS dataset’s size and check how robust the model is. The five-class training and validation accuracy for the expanded dataset is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.21\%$$\end{document}99.21% and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$75.50\%$$\end{document}75.50%. The two-class training and validation accuracy on augmented APTOS is \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.96\%$$\end{document}99.96% and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$97.05\%$$\end{document}97.05%. Extending the two-class model for the mixed dataset, we get a training and validation accuracy of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$99.92\%$$\end{document}99.92% and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$91.43\%$$\end{document}91.43%, respectively. We have also performed cross-dataset and mixed dataset testing to demonstrate the efficiency of DRISTI.

The present biometric market segment has been captured by compact, lightweight sensors which are capable of reading the biometric fluctuations of a user in real-time. This biometric market segment has further facilitated rise of a new ecosystem of wearable devices helpful in tracking the real-time physiological data for Healthcare-related analysis. However, the devices in the smart-wearable ecosystem are limited to capturing and displaying the biometrics without any prescriptive analytics. This paper addresses this gap to analyse the human emotion space based on an individual’s state of mind over the past 60 min and employs Deep Learning and Bayesian prediction techniques to predict the possibility of an impulsive outburst within upcoming few minutes. A lightweight smart processing device mounted with sensors captures the biometrics of the user and calibrate the same to the mental state of the user on a scale of zero to hundred. The results reveal that the deep learning algorithm along with the Bayesian probability module can predict the future mood fluctuations of the user with lower error than the other contemporary models. The predicted mood fluctuations has matched with the actual mood changes of the experimental subject within \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \pm 10 $$\end{document}±10 min of the predicted time index in 93% of the cases and within \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \pm 5 $$\end{document}±5 min in 82% of the cases.

The numerical nature of financial markets makes market forecasting and portfolio construction a good use case for machine learning (ML), a branch of artificial intelligence (AI). Over the past two decades, a number of academics worldwide (mostly from the field of computer science) produced a sizeable body of experimental research. Many publications claim highly accurate forecasts or highly profitable investment strategies. At the same time, the picture of real-world AI-driven investments is ambiguous and conspicuously lacking in high-profile success cases (while it is not lacking in high-profile failures). We conducted a literature review of 27 academic experiments spanning over two decades and contrasted them with real-life examples of machine learning-driven funds to try to explain this apparent contradiction. The specific contributions our article will make are as follows: (1) A comprehensive, thematic review (quantitative and qualitative) of multiple academic experiments from the investment management perspective. (2) A critical evaluation of running multiple versions of the same models in parallel and disclosing the best-performing ones only (“cherry-picking”). (3) Recommendations on how to approach future experiments so that their outcomes are unambiguously measurable and useful for the investment industry. (4) An in-depth comparison of real-life cases of ML-driven funds versus academic experiments. We will discuss whether present-day ML algorithms could make feasible and profitable investments in the equity markets.

With the recent spread of the SARS-CoV-2 virus, computer-aided diagnosis (CAD) has received more attention. The most important CAD application is to detect and classify pneumonia diseases using X-ray images, especially, in a critical period as pandemic of covid-19 that is kind of pneumonia. In this work, we aim to evaluate the performance of single and ensemble learning models for the pneumonia disease classification. The ensembles used are mainly based on fined-tuned versions of (InceptionResNet_V2, ResNet50 and MobileNet_V2). We collected a new dataset containing 6087 chest X-ray images in which we conduct comprehensive experiments. As a result, for a single model, we found out that InceptionResNet_V2 gives 93.52% of F1 score. In addition, ensemble of 3 models (ResNet50 with MobileNet_V2 with InceptionResNet_V2) shows more accurate than other ensembles constructed (94.84% of F1 score).

Accurate detection of hate speech against politicians, policy making and political ideas is crucial to maintain democracy and free speech. Unfortunately, the amount of labelled data necessary for training models to detect hate speech are limited and domain-dependent. In this paper, we address the issue of classification of hate speech against policy makers from Twitter in Italian, producing the first resource of this type in this language. We collected and annotated 1264 tweets, examined the cases of disagreements between annotators, and performed in-domain and cross-domain hate speech classifications with different features and algorithms. We achieved a performance of ROC AUC 0.83 and analyzed the most predictive attributes, also finding the different language features in the anti-policymakers and anti-immigration domains. Finally, we visualized networks of hashtags to capture the topics used in hateful and normal tweets., to appear in SN social sciences - special issue on hate speech

Application specific instruction-set processors (ASIP) are a programmable and flexible alternative of traditional finite state machine (FSM) controlled register-transfer level (RTL) designs for multimode basedband systems. In this paper, we present two ASIPs for small scale multiple-input multiple-output (MIMO) wireless communication systems that demonstrate the soundness and effectiveness of ASIPs for this type of applications. The first ASIP is programmed with multiple MIMO symbol detection algorithms for 4 × 4 systems. The supported detection algorithms are minimum mean-square error (MMSE), two variants of the selective spanning with fast enumeration (SSFE) and K-best list sphere detection (LSD). The second ASIP supports MMSE and zero-forcing dirty paper coding (ZF-DPC) algorithms for a base station (BS) with 4 antennas and for 4 users. Both ASIPs are based on transport triggered architecture (TTA) and are programmed with a retargetable compiler with high level language to meet the time-to-market requirements. The detection and precoding algorithms can be switched in the respective ASIPs based on the error-rate requirements. Depending on the algorithms, MIMO detection ASIP delivers 6.16—66.66 Mbps throughput at a clock frequency of 200 MHz on 90 nm technology. The precoder ASIP provides a throughput of 52.17 and 51.95 Mbps for MMSE and ZF-DPC precoding respectively at a clock frequency of 210 MHz on 90 nm technology.

The scarcity of fully-annotated data becomes the biggest obstacle that prevents many deep learning approaches from widely applied. Weakly-supervised visual learning which can utilize inexact annotations is developed rapidly to remedy such a situation. In this paper, we study the weakly-supervised task achieving pixel-level semantic segmentation only with image-level labels as supervision. Different from other methods, our approach tries to transform the weakly-supervised visual learning problem into a semi-supervised visual learning problem and then utilizes semi-supervised learning methods to solve it. Utilizing this transformation, we can adopt effective semi-supervised methods to perform transductive learning with context information. In the semi-supervised learning module, we propose to use the graph cut algorithm to label more supervision from the activation seeds generated from a classification network. The generated labels can provide the segmentation model with effective supervision information; moreover, the graph cut module can benefit from features extracted by the segmentation model. Then, each of them updates and optimizes the other iteratively until convergence. Experiment results on PASCAL VOC and COCO benchmarks demonstrate the effectiveness of the proposed deep graph cut algorithm for weakly-supervised semantic segmentation.

Satellite imagery is changing the way we understand and predict economic activity in the world. Advancements in satellite hardware and low-cost rocket launches have enabled near-real-time, high-resolution images covering the entire Earth. It is too labour-intensive, time-consuming and expensive for human annotators to analyse petabytes of satellite imagery manually. Current computer vision research exploring this problem still lack accuracy and prediction speed, both significantly important metrics for latency-sensitive automatized industrial applications. Here we address both of these challenges by proposing a set of improvements to the object recognition model design, training and complexity regularisation, applicable to a range of neural networks. Furthermore, we propose a fully convolutional neural network (FCN) architecture optimised for accurate and accelerated object recognition in multispectral satellite imagery. We show that our FCN exceeds human-level performance with state-of-the-art 97.67% accuracy over multiple sensors, it is able to generalize across dispersed scenery and outperforms other proposed methods to date. Its computationally light architecture delivers a fivefold improvement in training time and a rapid prediction, essential to real-time applications. To illustrate practical model effectiveness, we analyse it in algorithmic trading environment. Additionally, we publish a proprietary annotated satellite imagery dataset for further development in this research field. Our findings can be readily implemented for other real-time applications too.

In this paper, we propose to advance and transform today's healthcare system using a model of networked health care called Cybercare. Cybercare means "health care in cyberspace" - for example, doctors consulting with patients via videoconferencing across a distributed network; or patients receiving care locally - in neighborhoods, "minute clinics," and homes - using information technologies such as telemedicine, smartphones, and wearable sensors to link to tertiary medical specialists. This model contrasts with traditional health care, in which patients travel (often a great distance) to receive care from providers in a central hospital. The Cybercare model shifts health care provision from hospital to home; from specialist to generalist; and from treatment to prevention. Cybercare employs advanced technology to deliver services efficiently across the distributed network - for example, using telemedicine, wearable sensors and cell phones to link patients to specialists and upload their medical data in near-real time; using information technology (IT) to rapidly detect, track, and contain the spread of a global pandemic; or using cell phones to manage medical care in a disaster situation. Cybercare uses seven "pillars" of technology to provide medical care: genomics; telemedicine; robotics; simulation, including virtual and augmented reality; artificial intelligence (AI), including intelligent agents; the electronic medical record (EMR); and smartphones. All these technologies are evolving and blending. The technologies are integrated functionally because they underlie the Cybercare network, and/or form part of the care for patients using that distributed network. Moving health care provision to a networked, distributed model will save money, improve outcomes, facilitate access, improve security, increase patient and provider satisfaction, and may mitigate the international global burden of disease. In this paper we discuss how Cybercare is being implemented now, and envision its growth by 2030.

In this paper, we present an approach for automatically creating a CombinatoryCategorial Grammar (CCG) treebank from a dependency treebank for the Subject-Object-Verb language Hindi. Rather than a direct conversion from dependency trees to CCG trees, we propose a two stage approach: a language independent generic algorithm first extracts a CCG lexicon from the dependency treebank. An exhaustive CCG parser then creates a treebank of CCG derivations. We also discuss special cases of this generic algorithm to handle linguistic phenomena specific to Hindi. In doing so we extract different constructions with long-range dependencies like coordinate constructions and non-projective dependencies resulting from constructions like relative clauses, noun elaboration and verbal modifiers.

The high frequency data in intensive care unit is flashed on a screen for a few seconds and never used again. However, this data can be used by machine learning and data mining techniques to predict patient outcomes. Learning finite-state transducers (FSTs) have been widely used in problems where sequences need to be manipulated and insertions, deletions and substitutions need to be modeled. In this paper, we learned the edit distance costs of a symbolic univariate time series representation through a stochastic finite-state transducer to predict patient outcomes in intensive care units. The Nearest-Neighbor method with these learned costs was used to classify the patient status within an hour after 10 h of data. Several experiments were developed to estimate the parameters that better fit the model regarding the prediction metrics. Our best results are compared with published works, where most of the metrics (i.e., Accuracy, Precision and F-measure) were improved.

The 1993 paper in IJAIED on evaluation methods for Intelligent Tutoring Systems (ITS) still holds up well today. Basic evaluation techniques described in that paper remain in use. Approaches such as kappa scores, simulated learners and learning curves are refinements on past evaluation techniques. New approaches have also arisen, in part because of increases in the speed, storage capacity and connectivity of com- puters over the past 20 years. This has made possible techniques in crowd sourcing, propensity-score matching, educational data mining and learning analytics. This short paper outlines some of these approaches.

Usability tests are a part of user-centered design. Usability testing with disabled people is necessary, if they are among the potential users. Several researchers have already investigated usability methods with sighted people. However, research with blind users is insufficient, for example, due to different knowledge on the use of assistive technologies and the ability to analyze usability issues from inspection of non-visual output of assistive devices. From here, the authors aspire to extend theory and practice by investigating four usability methods involving the blind, visually impaired and sighted people. These usability methods comprise of local test, synchronous remote test, tactile paper prototyping and computer-based prototyping. In terms of effectiveness of evaluation and the experience of participants and the facilitator, local tests were compared with synchronous remote tests and tactile paper prototyping with computer-based prototyping. Through the comparison of local and synchronous remote tests, it has been found that the number of usability problems uncovered in different categories with both approaches was comparable. In terms of task completion time, there is a significant difference for blind participants, but not for the visually impaired and sighted. Most of the blind and visually impaired participants prefer the local test. As for the comparison of tactile paper prototyping and computer-based prototyping, it has been revealed that tactile paper prototyping provides a better overview of an application while the interaction with computer-based prototypes is closer to reality. Problems regarding the planning and conducting of these methods as they arise in particular with blind people were also discussed. Based on the authors' experiences, recommendations were provided for dealing with these problems from both the technical and the organization perspectives.

VerbNet—the most extensive online verb lexicon currently available for English—has proved useful in supporting a variety of NLP tasks. However, its exploitation in multilingual NLP has been limited by the fact that such classifications are available for few languages only. Since manual development of VerbNet is a major undertaking, researchers have recently translated VerbNet classes from English to other languages. However, no systematic investigation has been conducted into the applicability and accuracy of such a translation approach across different, typologically diverse languages. Our study is aimed at filling this gap. We develop a systematic method for translation of VerbNet classes from English to other languages which we first apply to Polish and subsequently to Croatian, Mandarin, Japanese, Italian, and Finnish. Our results on Polish demonstrate high translatability with all the classes (96% of English member verbs successfully translated into Polish) and strong inter-annotator agreement, revealing a promising degree of overlap in the resultant classifications. The results on other languages are equally promising. This demonstrates that VerbNet classes have strong cross-lingual potential and the proposed method could be applied to obtain gold standards for automatic verb classification in different languages. We make our annotation guidelines and the six language-specific verb classifications available with this paper. Electronic supplementary material The online version of this article (doi:10.1007/s10579-017-9403-x) contains supplementary material, which is available to authorized users.

High quality health services for psychiatric emergencies are difficult to maintain 24 h a day in sparsely populated areas, where the availability of specialists are limited. To overcome this challenge, the University Hospital of North Norway implemented an on-call system in psychiatric emergencies, by which psychiatrists are accessible 24/7 for telephone and videoconferencing consultations with patients and nurses at three regional psychiatric centers. The purpose of this study was to explore the usefulness of videoconferencing consultations in psychiatric emergencies. The on-call psychiatrists and nurses at the regional psychiatric centers who participated in at least one videoconferencing consultation were recruited for semi-structured interviews: 5 psychiatrists and 19 nurses for a total of 24 participants. The interviews were performed from July of 2012 to June of 2013. The interviews were transcribed and analyzed using a stepwise-deductive-inductive approach supported by the HyperRESEARCH analysis software. After the introduction of the videoconferencing system, telephone consultations were still used for straightforward situations when further treatment or admission decisions were obvious. Videoconferencing consultations, during which patients were present, were useful for challenging situations when there was 1) uncertainty regarding the degree of illness or the level of treatment, 2) a need to clarify the severity of the patient condition, 3) a need to build an alliance with the patient, and 4) disagreement either between health personnel or between the patient and health personnel. Videoconferencing is useful for challenging and complex psychiatric emergencies and is a suitable tool for building high quality, decentralized psychiatric services.

The capability of UML profiles to serve as annotation mechanism has been recognized in both research and industry. Today’s modeling tools offer profiles specific to platforms, such as Java, as they facilitate model-based engineering approaches. However, considering the large number of possible annotations in Java, manually developing the corresponding profiles would only be achievable by huge development and maintenance efforts. Thus, leveraging annotation-based modeling requires an automated approach capable of generating platform-specific profiles from Java libraries. To address this challenge, we present the fully automated transformation chain realized by Jump, thereby continuing existing mapping efforts between Java and UML by emphasizing on annotations and profiles. The evaluation of Jump shows that it scales for large Java libraries and generates profiles of equal or even improved quality compared to profiles currently used in practice. Furthermore, we demonstrate the practical value of Jump by contributing profiles that facilitate reverse engineering and forward engineering processes for the Java platform by applying it to a modernization scenario. published

Fast changing hardware and software technologies in addition to larger and more specialized customer bases demand software tailored to meet very diverse requirements. Software development approaches that aim at capturing this diversity on a single consolidated platform (i.e. software product lines [5]) often require large upfront investments of time and money. Alternatively, companies resort to developing one variant of a software product at a time by reusing as much as possible from already existing product variants (i.e. clone-and-own [2]). However, identifying and extracting the parts to reuse is an error-prone and inefficient task. Hence, more disciplined and systematic approaches are needed to cope with the complexity of developing and maintaining sets of product variants. Such approaches require detailed information about the product variants, the features they provide and their relations.Variability is the capacity of software to vary [6]. Its effective management requires variability information such as the set of possible product variants, the features they provide, how they are related, and how they are implemented. In this work we present an approach for extracting variability information from sets of related product variants. We compute traces from features and feature interactions to their implementation artifacts and dependencies between features. A Trace is a link between a source and a target artifact (in this context between features and artifacts) [1]. Traceability is the potential for traces to be established and used. Variability management is paramount for coping with scenarios where multiple product variants must be developed and maintained such as:Supporting and Enhancing Clone-and-Own Reuse. Clone-and-own is a manual ad hoc software reuse approach where new product variants are created by reusing parts from already existing variants [2]. The parts to be reused must first be located in the existing variants, then extracted, merged, and completed to obtain the new working variant. This process is repeated for each new variant required. Variability information helps to locate reusable features and their implementing artifacts. It even makes it possible to partially automate reuse [3].Reverse Engineering Software Product Lines (SPLs). An SPL consists of a set of related product variants that are distinguished by the set of features they provide [5]. In cases where the problem domain and the existing set of variants have become mature and stable, software companies can consider to consolidate existing variants into a fully integrated SPL platform. Variability information is not only highly useful but even necessary for reverse engineering sets of variants into an SPL [3].Extending an SPL. In cases where an SPL already exists, the need of providing new features to meet new customer requirements may still arise. In such cases product lines have to be extended to provide such new features, a process that requires knowledge about already existing features and how they are related and implemented.

The establishment of secure secret keys ahead of transmissions is one of the key issues in the field of information security. The security of traditional cryptographic secret key establishment mechanisms is seriously challenged by computing-intensive attacks, with the fast growth of high-performance computing. As an alternative, considerable efforts have been made to develop physical (PHY) layer security measures in recent years, such as link-signature-based (LSB) secret key extraction techniques. Those mechanisms have been believed secure, based on the fundamental assumption that wireless signals received at two locations are uncorrelated when separated by more than half a wavelength. However, this assumption does not hold in some circumstances under latest observations, rendering LSB key extraction mechanisms vulnerable to attacks. To address this problem, the formal theoretical analysis on channel correlations in both real indoor and outdoor environments is provided in this paper. Moreover, this paper proposes empirical statistical inference attacks (SIA) against LSB key extraction, whereby an adversary infers the signature of a target link. Consequently, the secret key extracted from that signature has been recovered by observing the surrounding links. In contrast to prior literature that assumes theoretical link-correlation models for the inference, our study does not make any assumption on link correlation. Instead, we employ machine learning (ML) methods for link inference based on empirically measured link signatures. We further propose a countermeasure against the SIAs, called forward-backward cooperative key extraction protocol with helpers (FBCH). In the FBCH, helpers (other trusted wireless nodes) are introduced to provide more randomness in the key extraction. Our experimental results have shown that the proposed inference methods are still quite effective even without making assumptions on link correlation. Furthermore, the effectiveness of the proposed FBCH protocol is validated by our experiment results.

In today’s design — manufacturing context, designers often modify existing 3D shapes (or design models) instead of creating a new design from scratch. This requires the ability to search an existing database of designs/3D models to identify and extract similar designs. Shape Retrieval Tools (SRTs) have been developed to provide an essential role in saving time and effort to retrieve and generate new designs. The capabilities of commercially available SRTs vary based on the form of the input design model, the search technique or algorithm used, the search/retrieval time, ease of use, and the quality of results. The focus of this paper is to study of their capabilities, performances, and differences and develop criteria to compare the effectiveness and performance of such Shape Retrieval Tools. Current search evaluation methods, such as precision and recall, are based on human interpretation of the results. This paper presents a holistic set of metrics for comparing the performance and effectiveness of SRTs, including data input options (to search), effectiveness of the search process, the associated retrieval time, overall ease of use, and additional data retrieval details. An algorithm is proposed to objectively analyze the search results based on the proposed Model Match Ratio (MMR), computed by the variance between the input and retrieved geometries. The search results are usually presented in a rank order list. A Precision Sequence Metric (PSM) is developed to evaluate the retrieved list by ranking the retrieved results based on the MMR for evaluating the quality of the search. The proposed evaluation algorithm was tested on several design models (and their subsequent retrieval results) involving three SRTs (Vizseek, Geolus, and CADENAS); the results of the comparison of the performance of these SRTs are discussed in this paper.

Deep learning (DL) is gaining significant prevalence in every field of study due to its domination in training large data sets. However, several applications are utilizing machine learning (ML) methods from the past several years and reported good performance. However, their limitations in terms of data complexity give rise to DL methods. Intrusion detection is one of the prominent areas in which researchers are extending DL methods. Even though several excellent surveys cover the growing body of research on this subject, the literature lacks a detailed comparison of ML methods such as ANN, SVM, fuzzy approach, swarm intelligence and evolutionary computation methods in intrusion detection, particularly on recent research. In this context, the present paper deals with the systematic review of ML methods and DL methods in intrusion detection. In addition to reviewing ML and DL methods, this paper also focuses on benchmark datasets, performance evaluation measures and various applications of DL methods for intrusion detection. The present paper summarizes the recent work, compares their experimental results for detecting network intrusions. Furthermore, current research challenges are identified for helping fellow researchers in the era of DL-based intrusion detection.

This paper investigates the art design of user interaction experience in mobile systems through the methods of contextual perception and machine learning. The theoretical foundations for the design of intangible cultural heritage interactive display resources include digital display theory, intangible cultural heritage education theory, embodied cognition theory, and gamification design theory. Based on the modelling and analysis of the theoretical foundations, the design principles are derived, including the heritage education principle, the somatic interaction design principle, and the content design principle. In this paper, we use ontologies to construct user knowledge models, fuse multi-situational similarity metrics, screen out candidate neighbour sets through preliminary screening, then combine the user's activity to construct time-based weight tensor scores, use tensor decomposition to obtain recommendation evaluation values, and finally use the recommendation evaluation values to make artistic recommendations. The experimental results show that the algorithm can still obtain a good recommendation implementation in the case of extremely sparse data. The analysis of the fit between augmented reality and folk art appreciation class, as well as the reference to relevant application cases, make design and practice of augmented reality application in folk art appreciation class, try to solve the common problems in folk art appreciation class, analyse the feedback effect and make a summary and outlook.

Competitiveness, defined as the rate of success in attracting and maintaining industries to foster the sustained improvement in citizens’ wellbeing, has been a long-pursued goal for regions and nations. Today’s rapid advancements in technology, especially in telecommunications, open challenges for decision and policy makers to generate effective and efficient solutions in a global scenario. In this context, the latest developments in artificial intelligence, machine learning and deep learning open new paths for describing, analyzing, and representing complex phenomena in systemic environments. This paper presents a model using a neural network to predict the behavior of competitive benchmarks using public expenditure variables. The theory of control, in which the neural network approach is based, offers some advantages such as solving the problem while considering the dynamic nature of the phenomenon and allowing control blocks to be implemented in a straightforward method. The present paper establishes a neural network model that links control, administration, and systems theories in a statistically sound approach that connects both sets of variables, opening the path for extensions that allow optimal allocation of resources.

CAPTCHA stands for Completely Automated Public Turing Test to Tell Computers and Human Apart. CAPTCHA is used for internet security. A few CAPTCHA schemes are available today like, text-based, audio-based, video/animation-based, puzzle based etc. In this paper, all these types are collaborating at single place to analyze. The main aim of this article is to present a literature to identify and recognize CAPTCHA, its types, the creation and breaking techniques. It is a systematic and complete analysis of all available CAPTCHA types. In this paper, 16 text-based CAPTCHA’s generation methods are discussed with usability and security ranges from 3 to 100 and 65 to 100%, respectively. The security and usability measures are not calculated/sustained using some known English schemes. Out of 16 reviewed CAPTCHAs, 12 are based on English language, 1 on Arabic language, 1 on Chinese language, 1 on Devanagari language and 1 on Gurumukhi script. The designs are made segment proof with overlapping random shapes, overlapping characters, clasping, different colors and different shades. For making recognition proof many techniques are used like image masking, local and global warping; broken characters, random rotation, arcs, jaws, etc. Approximately 50 schemes, especially based on the English language, are successfully broken with a success rate that ranges from 2 to 100%. The techniques that are used to break these schemes include shape context matching, distortion estimation, Log Gabor 2D filter, horizontal and vertical projection (for a segment the letters) are used. For recognition CNN, KNN, DNN and MCDNN are used. Almost 15 images-based CAPTCHAs are discussed that are designed with usability and security range 90–100 and 17–100%, respectively. Out of these 5 schemes are successfully broken with a success rate ranging between 7 and 100%. The K-NN and SVM are mostly used algorithms to recognize the images. Audio based CAPTCHAs (5 designs) are discussed with usability and security range from 68.5 to 100 and 100%, respectively. The broken rate of these audio schemes is also 45–75%. These schemes are broken with SVM and K-NN algorithms. The paper also discusses 4 popular video-based designs that provide usability and security that ranges from 75 to 100 and 98 to 100, respectively. These schemes are also compromised with broken rate 16–10% using SIFT, NN and simple OCR techniques. The paper can be a benchmark to precede any specific research to dive into any one of these types.

Sentiment analysis is now a prominent field of interest owing to a growing trend of users expressing their opinions on social media, review pages, feedback forms, and other online channels. The machine learning approach to sentiment analysis focuses on feature extraction methods like constructing lexicons to learn sentiment polarity or learning word embeddings and applying them for their use in machine learning algorithms for sentiment classification. But most popular machine learning approaches still cannot capture nuanced emotions like sarcasm, irony, etc. Emojis are now being used along with text by the users to express emotions and hence can help researchers improve sentiment classification tasks. Sentiment analysis powered by emojis is still in the nascent phase and has gained some pace in the last five years. The primary goal of this paper is to discuss the use of emojis that supplement the text to express different emotions. This paper compares some traditional text-based word embeddings and lexicons. Then the paper discusses the evolution of emoji-based lexicons and emoji embeddings. Further, some deep learning approaches using emojis to improve existing sentiment classification tasks are studied. The main contribution of this paper is to survey various approaches to use emojis in sentiment analysis which to the best of our knowledge has not been done till now.

Executable files still remain popular to compromise the endpoint computers. These executable files are often obfuscated to avoid anti-virus programs. To examine all suspicious files from the Internet, dynamic analysis requires too much time. Therefore, a fast filtering method is required. With the recent development of natural language processing (NLP) techniques, printable strings became more effective to detect malware. The combination of the printable strings and NLP techniques can be used as a filtering method. In this paper, we apply NLP techniques to malware detection. This paper reveals that printable strings with NLP techniques are effective for detecting malware in a practical environment. Our dataset consists of more than 500,000 samples obtained from multiple sources. Our experimental results demonstrate that our method is effective to not only subspecies of the existing malware, but also new malware. Our method is effective against packed malware and anti-debugging techniques.

From exchanging budgetary instruments to tracking individual spending plans to detail a business's profit, money-related organisations utilise computational innovation day by day. Here in this paper, we focus on the significance of innovation in accounts such as financial risk management and stock prediction. We discuss two significant algorithms that have a notable role in stock forecasting. Artificial Neural Networks (ANN), as absenteeism of some data points, does not hamper the network functioning. Secondly, Support Vector Machines (SVM) has several features, and due to simple decision boundaries, it avoids over-fitting. The paper first looks at the different technologies applied in stock market prediction. It examines how sentimental analysis, decision trees, moving average algorithm, and data mining is applied in various stock prediction scenarios. The paper covers the recent past studies to explore the concepts and methodologies through which ANN's and SVM's have been used. Additionally, the paper incorporates significant aspects of novel methods and technologies in which ANN as a hybrid model like ANN-MLP, GARCH-MLP, a combination of the Backpropagation algorithm and Multilayer Feed-forward network, yields better results. Simultaneously, SVM's have been successfully applied in stock prediction, giving an accuracy of about 60%–70% for simple SVM, which is further improved by combining methods like Random Forest, Genetic Algorithm more accurate outcomes. Further, we present our thoughts on where SVM's and ANN's stand as prediction algorithms and challenges like the time constraint, current scenarios, data limitation, and cold start problems were raised. Conclusively SVM and ANN played prominent roles in tackling these issue to an extent and can further be enhanced with their integration with other novel techniques resulting in hybrid methodologies. It will lead students, researchers and financial enthusiasts to more potent approaches for Stock forecasting.

When the turning tool has worn and failed but the failure is not found, if it continues to be used for processing, it will break, and cause the workpiece to be scrapped, and even damage the machine tool. In order to avoid the loss caused by turning tool wear, the remaining useful life (RUL) prediction of turning tool wear has become a hot research topic in recent years. For RUL prediction in turning tools, the traditional machine is difficult to acquire sufficient degradation data and inconsistent data distribution among different turning tools in engineering, and they cannot provide better prediction accuracy to some extent. To solve the above problems, this paper proposes a multi-granularity feature extraction (MGFE) method based on the gray-level co-occurrence matrix (GLCM) and random forest (RF). Moreover, a health indicator (HI) of turning tools in the source domain was obtained. The common representative features in HI sequence of target domain was transferred to source domain and builds the condition monitoring and life prediction system of turning tools based on extreme learning machine and transfer learning. Finally, extreme vector machine (ELM) is used to construct the RUL prediction model. The research results show that the model constructed in this paper is effective in RUL prediction and can significantly improve the prediction accuracy of remaining useful life.

Rockburst is a typical dynamic phenomenon of mine destruction. With the expansion of mining scale and mining depth, its harm is becoming more and more serious. This has become an important problem to be solved urgently in the mining industry. This paper aims to study the prediction and prediction of rockburst risk based on particle swarm algorithm and neural network. This paper proposes a shock risk assessment method based on BP neural network. It uses existing shock pressure data to establish a regression model through BP network, and uses PSO algorithm to optimize connection weights to evaluate the slow convergence of BP network and easy to fall into local optimality. The shortcomings of and the degree of convergence was evaluated. In addition, this paper proposes a rockburst risk prediction method. In order to improve the accuracy of rockburst prediction, regional prevention and control measures and certain risk mitigation methods can also be quickly adopted. It not only identifies the mechanical properties of coal and rock mass in the laboratory, but also judges the possibility of rockbursts during mining. The experimental results in this paper show that 10 main factors that affect rockbursts are selected, 20 standard mechanical data sets are used, and a PSO-BP-based mine explosion risk assessment model is established, and the model is compared with the standard BP model Make a comparison. The results show that compared with the standard BP model, the evaluation accuracy of the PSO-BP model is increased by 15%. Finally, an example of mine risk assessment verifies the accuracy and overall applicability of the method.

Extreme learning machine (ELM) is a training algorithm for single hidden layer feedforward neural network (SLFN), which converges much faster than traditional methods and yields promising performance. In this paper, we hope to present a comprehensive review on ELM. Firstly, we will focus on the theoretical analysis including universal approximation theory and generalization. Then, the various improvements are listed, which help ELM works better in terms of stability, efficiency, and accuracy. Because of its outstanding performance, ELM has been successfully applied in many real-time learning tasks for classification, clustering, and regression. Besides, we report the applications of ELM in medical imaging: MRI, CT, and mammogram. The controversies of ELM were also discussed in this paper. We aim to report these advances and find some future perspectives.

Data protection across organizations is limiting the application of centralized learning (CL) techniques. Federated learning (FL) enables multiple participants to build a learning model without sharing data. Nevertheless, there is very few research works on FL in intelligent manufacturing. This paper presents the results of an empirical study on failure prediction in the production line based on FL. This paper (1) designs Federated Support Vector Machine and federated random forest algorithms for the horizontal FL and vertical FL scenarios, respectively; (2) proposes an experiment process for evaluating the effectiveness between the FL and CL algorithms; (3) finds that the performance of FL and CL are not significantly different on the global testing data, on the random partial testing data, and on the estimated unknown Bosch data, respectively. The fact that the testing data is heterogeneous enhances our findings. Our study reveals that FL can replace CL for failure prediction.

Power system protection and asset management have drawn the attention of researchers for several decades; but they still suffer from unresolved and challenging technical issues. The situation has been recently exacerbated in the wake of the ever-changing landscape of power systems driven by the growing uncertainty and volatility subsequent to the vast renewable energy integration, more frequent natural extreme events due to climate changes, increasing malicious cyberattacks, and more constrained transmission systems as the result of load growth and limited investments. On the opposite side, the proliferation of advanced measuring devices such as phasor measurement units, emerging electric and non-electric sensors, and Internet of Thing (IoT)-enabled data gathering platforms continually expand/nourish the databases; they hence offer unprecedented opportunities to take the advantage of data-driven techniques. Machine learning (ML) as a principal class of artificial intelligence is the perfect match solution to this need and has newly revoked many researchers’ interests to tackle the problems excluding their exact/detailed models. This paper aims to provide an overview on applications of ML techniques in power system protection and asset management. This paper elaborates on issues pertaining to (1) synchronous generators, (2) power transformers, (3) transmission lines, and (4) special and system-integrity protection schemes. In addition to the opportunities offered by the ML techniques, this paper discourses on the barriers and challenges to the wide-spread application of ML techniques in real-world practices.

The new energy vehicle system is in the initial stage of application, so the probability of fault is greater. Therefore, its reliability urgently needs to be improved. In order to improve the fault diagnosis effect of new energy vehicles, this paper proposes a fault diagnosis system of new energy vehicle electric drive system based on improved machine learning and also proposes several typical fault detection and diagnosis methods. Through the study of the operating characteristics and structural characteristics of the electric drive system and the analysis of the fault mechanism, this paper classifies the main faults of the electric drive system according to the self-test requirements of the electric drive system and the online diagnosis requirements and presents the self-test fault tree and online diagnosis fault tree of the electric drive system. In addition, this paper designs experiments to verify the performance of the system proposed by this paper, simulates the current common system fault conditions, and uses the system constructed in this paper to perform system fault diagnosis. The research results show that the performance of the fault diagnosis system for drive energy vehicles constructed in this paper is reliable.

At present, the mortgage loan of farmland contractual management right is in the pilot and exploration stage. In this overall environment, how to formulate an appropriate pricing model according to the particularity of farmland property right and the actual situation is the theoretical support to be solved. On the basis of theoretical analysis, model construction and analysis of existing problems, this paper establishes a binary tree pricing method of farmland management right mortgage based on machine learning and complex network algorithm. This paper analyzes the basic theory of rural land management right mortgage pricing and constructs the pricing model. Based on the Markov decision process model of dynamic pricing problem, this paper proposes the solution algorithm and proposes the dynamic pricing solution for this paper from two aspects of value function and strategy gradient reinforcement learning algorithm. This paper analyzes the basic principle of A3C algorithm and expounds the algorithm design and neural network architecture for the specific dynamic pricing problem in this paper. The dynamic pricing method of rural land mortgage based on deep reinforcement learning designed in this paper can well predict the real price of rural land management right mortgage pricing and effectively optimize the mortgage loan environment of rural contracted land management right, which is the innovation of financial system in the current new rural construction.

Stock market is a dynamic and volatile market that is considered as time series data. The growth of financial data exposed the computational efficiency of the conventional systems. This paper proposed a hybrid deep learning model based on Long Short- Term Memory (LSTM) and Artificial Bee Colony (ABC) algorithm. ABC is best fit for hyper parameter selection for deep LSTM models and maintains the equilibrium of exploitation and exploration issues. Handling a large volume of multidimensional reviews from social media is a major challenge. This paper evolves the multiple aspects of market sentiments and uses the reliable Big data platform Hadoop ecosystem and its services to compute sentiment polarity index. The ABC-LSTM hybrid model is validated with other core and hybrid models with evolutionary algorithms as Differential Evolution (DE) and Genetic Algorithm (GA). For the experiments, 10 years of historical datasets and social media reviews of IT sector funds Apple Inc. (AAPL), Microsoft corporation (MSFT) and Intel corporation (INTL) from NASDAQ GS, an American stock exchange are considered to validate hybrid forecasting models. Proposed algorithm ABC-LSTM is used to tune the hyperparameters (window size, LSTM units, dropout probability, epochs, batch size and learning rate) and evaluated through Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE) as loss function. Performance analysis proves that with sentiment polarity, ABC optimized LSTM obtains improved forecasting accuracy over its counterpart models.

The Internet of Things is a rapidly evolving technology in which interconnected computing devices and sensors share data over the network to decipher different problems and deliver new services. For example, IoT is the key enabling technology for smart homes. Smart home technology provides many facilities to users like temperature monitoring, smoke detection, automatic light control, smart locks, etc. However, it also opens the door to new set of security and privacy issues, for example, the private data of users can be accessed by taking control over surveillance devices or activating false fire alarms, etc. These challenges make smart homes feeble to various types of security attacks and people are reluctant to adopt this technology due to the security issues. In this survey paper, we throw light on IoT, how IoT is growing, objects and their specifications, the layered structure of the IoT environment, and various security challenges for each layer that occur in the smart home. This paper not only presents the challenges and issues that emerge in IoT-based smart homes but also presents some solutions that would help to overcome these security challenges.

The rapid proliferation and progress of Wireless Sensor Networks (WSN) and Internet of Things (IoT) has conducted to the formation of a gigantic amount of data and a growing need to multiple new services and resources. In spite of the main role of Cloud computing in solving these issues, IoT applications need more reduced latency with mobility support and location awareness. To overcome the mentioned limits, a new concept favouring the integration of Fog computing onto IoT is increasingly utilized. It is a motivating scheme that offers a timely task execution and data management at the network edge, in a distributed way, with the collaboration of nearby nodes. In this paper, we provide a complete study of the integration of Fog onto IoT. We discuss the various challenges that are facing the Fog/IoT paradigm and specify the main contributions that have been proposed to overcome these challenges. We give also an insight on the relationship between IoT/Fog integration concept, and other leading technologies. The open issues that need more investigation are highlighted in this paper to identify clearly the research gaps in the area of Fog computing integration onto IoT.

This paper proposes a new way of Quantum signature masked authentication, in which both user and certifying authority is needed for generating final signature. Our scheme uses decoy particle to protect key from all the attacks in the transmission process and does not require encryption of each key. The paper also shows that our scheme is protected against impersonation, forgery and outside attack and is more efficient than the other masked schemes existing in quantum cryptography.

Food has a direct impact on an individual's life and is a significant area for the research community. Computational techniques in food-related computing are performed to address various food-related issues in the field of agronomy, medicine, biology etc. In this paper, we present a comprehensive review of research exclusively on segmentation techniques used for food computing. This paper illustrates the viable segmentation techniques used for food image segmentation. It also provides a comprehensive review of the same. A relevant survey on 66 research papers has been done to provide different food image segmentation techniques. A comparative study among these techniques is also done based on different parameters like type of algorithm, segmentation technique, dataset, and accuracy. Moreover, this paper focuses on research challenges in food recognition. Also a framework has been proposed in this paper to overcome the problem of watershed and OTSU algorithm.

OCR (Optical Character Recognition) for scanned paper invoices is very challenging due to the variability of 19 invoice layouts, different information fields, large data tables, and low scanning quality. In this case, table structure recognition is a critical task in which all rows, columns, and cells must be accurately positioned and extracted. Existing methods such as DeepDeSRT only dealt with high-quality born-digital images (e.g., PDF) with low noise and apparent table structure. This paper proposes an efficient method called CluSTi (Clustering method for recognition of the Structure of Tables in invoice scanned Images). The contributions of CluSTi are three-fold. Firstly, it removes heavy noises in the table images using a clustering algorithm. Secondly, it extracts all text boxes using state-of-the-art text recognition. Thirdly, based on the horizontal and vertical clustering algorithm with optimized parameters, CluSTi groups the text boxes into their correct rows and columns, respectively. The method was evaluated on three datasets: i) 397 public scanned images; ii) 193 PDF document images from ICDAR 2013 competition dataset; and iii) 281 PDF document images from ICDAR 2019’s numeric tables. The evaluation results showed that CluSTi achieved an F1-score of 87.5%, 98.5%, and 94.5%, respectively. Our method also outperformed DeepDeSRT with an F1-score of 91.44% on only 34 images from the ICDAR 2013 competition dataset. To the best of our knowledge, CluSTi is the first method to tackle the table structure recognition problem on scanned images.

A critical issue in data mining and machine learning is feature selection. The crucial part is how to specify the eminent problem-relevant features out of a collection of features contained in a dataset. Feature selection process goes with the pre processing steps in knowledge revelation (KDD process). It aids in eliminating the unnecessary (redundant) and unrelated (irrelevant) features in order to improve the fulfillment of classifying algorithms. It chooses the most optimal count of features that is best suited to classification model which in turn advance the learning process. As such, the correctness (accuracy) of classification increases. Thus, in this paper we have proposed a two-staged hybrid arrangement of model that contains filter-based approach in the first stage to filter out the unnecessary and unrelated features and then providing these acquired features as input to the next stage that is the wrapper method by availing the recent swarm based algorithm, namely, salp swarm algorithm or SSA. The proposed model is named as MRMR-SSA. The binary version of SSA is utilized to evaluate the features that can either take the feature as 1 or discard it as 0. Specific classifiers like XGBoost, AdaBoost, Random forests and Logistic Regression are made in use in this paper. Accuracy is considered to measure the performance of each classifier. An analogy is made for the proposed hybrid feature selection approach with a few familiar algorithms specifically MRMR-PSO, MRMR-GA, MRMR-ALO and MRMR-ACO. The proposed hybrid approach leaves behind other given hybrid methods.

Deep learning is increasingly studied in the prediction of cancer yet few deep learning systems have been introduced for daily use for such purpose. The manual scanning, reading, and analysis by radiologists to detect cancer are very time-consuming processes due to their large volume. Although many types of research have been conducted in this area, the use of their results in the diagnosis of prostate cancer is yet to be properly carried out. In this paper, a Data, Detection, and View (DDV) taxonomy is introduced that defines each major component, which is required to implement a proper deep learning prostate cancer detection system. The proposed taxonomy is a step toward developing a way to assist the pathologists for early detecting prostate cancer and hence facilitating the patients to seek speedy counseling from the doctors. If the diagnosis of cancer can be performed in the early stages then it can be prevented from spreading to other cells. The components of the proposed taxonomy must be reviewed and used as an evaluation and validation criteria for approving the deep learning classification model to be used in real-world clinical practices. Through the study of 22 state-of-the-art research papers in the field of deep learning-based prostate cancer classification system, we proved the effectiveness and robustness of the proposed DDV taxonomy system.

This research paper proposes a novel approach named priority-based load balancing (PLB) for cloud computing environment. The PLB provides a resilient and adaptive task scheduling using multi-queues. Numerous strategies have already been proposed in the past researches to prioritize the tasks and mapping all the tasks to different resources available on the cloud. There is still a hindrance in the performance due to the negligible attention paid to the unused resources and tasks having low priority, eventually leading to starvation problem. To this end, the PLB algorithm has been partitioned into four sub-procedures, namely (i) Starvation-free task allocation, (ii) Inserting tasks into the dispatcher, (iii) Re-ordering tasks inside the queues and eventually, (iv) Mapping tasks onto the Virtual Machines (VMs) calculating the cost incurred for all the corresponding VMs. The sole motivation of this research work is to optimize the performance parameters by allocating all the jobs to all the available resources in the workflow model. It also consolidates the job categorization in the priority-based multi-queues, while filtering tasks from all the queues to overcome the deprivation of low priority tasks. In this paper, a test-bed setup has been deployed using CloudSim 3 and TCS WAN emulator for experimentation and results evaluation. The experimental setup imbibes different aspects such as performance measures, average response time, makespan time in order to ascertain efficiency, resource utilization ratio and bandwidth of the workflow model. The obtained results are further compared with five different approaches including- First Come First Serve, Round Robin, Min–Min, Max–Min and ACO and it was observed that the proposed strategy yielded more efficiency and accuracy in most of the cases. The experimental results have been further validated and demonstrated in order to justify the claims of the proposed approach, being able to tackle out different priority tasks and resource allocation in a stable and optimum manner.

The construction of risk knowledge graphs aims at the effective organization and utilization of enterprise knowledge resources in big data environments. To address the problem of static mapping in existing enterprise knowledge graphs, this paper introduces the time dimension to describe the evolutionary characteristics of enterprise risk events, such as dynamics, suddenness and timeliness. Through information extraction, knowledge fusion, ontology construction and dynamic knowledge reasoning about risk knowledge, a bottom-up enterprise dynamic risk knowledge graph is systematically constructed. In the knowledge fusion link, aiming at the imbalanced classification problem for the entity samples of a data set, this paper proposes the ResNet dynamic knowledge reasoning method to improve the loss balance function of the Multi-Net model. The experiments show that the new model can effectively improve the accuracy of entity and relationship prediction. Finally, the knowledge graph is applied to an intelligent question-answering system.