Yanfei Li, Gang Liu, ZiwenTang and Hu wang, School of Computer science and technology, Xidian University, Xian, China
Differential privacy has become one of the most effective privacy protection methods because of its strong theoretical verification foundation and strict mathematical axioms. However, the existing differential privacy methods have high publishing errors when publishing infinite data streams, which reduces the accuracy of published data. Aiming at the feature that the data at adjacent moments in the infinite data stream are correlated, this paper proposes an infinite data stream publishing algorithm with low publishing errors and meets the requirements of differential privacy. Based on adjusting the threshold to meet the sampling rate, this method uses an adaptive budget allocation mechanism to protect the sampling point privacy and uses the Kalman filter mechanism to correct the non-sampling point data. Also, the test on data sets from two different fields shows that our algorithm can effectively reduce publishing errors and significantly improve the availability of published data.
Differential Privacy, Infinite data streams, Privacy preservation, Sampling.
Ouafae Elaeraj and Cherkaoui Leghris, L@M, RTM Team, Faculty of Sciences and Techniques Mohammedia, Hassan II University of Casablanca, Morocco
With the popularization of the Internet and local area networks, malicious attacks and intrusions into computer systems are increasing. The design and implementation of intrusion detection systems became extremely important to help maintain good network security. Support vector machines (SVM), a classic pattern recognition tool, has been widely used in intrusion detection. The mature and robust support vector machine algorithm is used and the grid search method is used for parameter optimization. This paper presents a new SVM model enriched with a Gaussian kernel function based on the features of the training data for intrusion detection. The new model is tested with the CICIDS2017 dataset. The test proves better results in terms of detection efficiency and false alarm rate, which can give better coverage and make the detection more effective.
Intrusion detection System, Support vector machines, Machine Learning.
Igli Tafa and Kevin Shahollari, Department of Computer Engineering, Polytechnic University of Tirana, Tirana
The rapid evolution of technology has increased the role of cybersecurity and put it at the center of national critical infrastructure. This role supports and guarantees the vital services of (CNI) while provides the proper functionalities for running operations between the public and private sectors. This evolution has had the same impact on cyberattack tools, methods, techniques used to gain unauthorized access to these computer systems that contain confidential and high-value information in the digital data sales market or as it called "darkweb". As a result, it has become necessary to monitor all events of the National Critical Infrastructure (CNI) computer systems. This proposed system uses a centralized National SIEM (N-SIEM) specializing in the correlation of security events caused by cyber attacks, collected by CNIs systems while integrating with an International Cyber Threat Intelligence system (ICTI-CNI). In addition, this conceptual model collects security breach events of CNIs systems, analyzes only cyber attacks, and correlates these security events in real-time with an intelligent automated platform while reducing the response time of security analysts.
CNI, N-SIEM, ICTI-CNI, IOC, cyber attacks security events.
Sphamandla May, Omowunmi Isafiade and Olasupo Ajayi, University of the Western Cape, South Africa
It is no secret that crime affects everyone and has devastating socio-economic impact with long lasting effects. For this reason, law enforcement agents work tirelessly to reduce crime. However, despite their best efforts, crime still prevails and is constantly on the rise. Predictive approaches have shown positive results in mitigating this problem, hence in this paper, we have adopted the well known Naive Bayes algorithm to tackle crime. In this work, we augmented the classic Naive Bayes with recursive feature elimination and applied it to crime prediction. When compared with the original Naive Bayes, it was observed that our approach indeed improved the performance of Naive Bayes by about 30%. We further bench-marked our Naive Bayes variant with other predictive algorithms such as Random Forest and Extremely Randomized Trees, and obtained results showed that our variant was equally as good and even better in some instances.
Crime prediction, Extremely Randomized Trees, Random Forest, Naive Bayes.
Sukanya Nath and Mascha Kurpicz-Briki, Bern University of Applied Sciences, Switzerland
Burnout is a major problem of todays society, in particular in crisis times such as a global pandemic situation. Burnout detection is hard, because the symptoms often overlap with other diseases and syndromes. Typical clinical approaches are using inventories to assess burnout for their patients, even though free-text approaches are considered promising. In research of natural language processing applied to mental health, often data from social media is used and not real patient data, which leads to some limitations for the application in clinical use cases. In this paper, we fill the gap and provide a dataset using extracts from interviews with burnout patients containing 216 records. We train an SVM classifier to detect burnout in text snippets with an accuracy of around 80%, which is clearly higher than the random baseline of our setup. This provides the foundation for a next generation of clinical methods based on NLP.
Natural Language Processing, Psychology, Burnout, Machine Learning.
Yew Kee Wong, School of Information Engineering, HuangHuai University, Henan, China
Artificial intelligence has been a buzz word that is impacting every industry in the world. With the rise of such advanced technology, there will be always a question regarding its impact on our social life, environment and economy thus impacting all efforts exerted towards sustainable development. In the information era, enormous amounts of data have become available on hand to decision makers. Big data refers to datasets that are not only big, but also high in variety and velocity, which makes them difficult to handle using traditional tools and techniques. Due to the rapid growth of such data, solutions need to be studied and provided in order to handle and extract value and knowledge from these datasets for different industries and business operations. Numerous use cases have shown that AI can ensure an effective supply of information to citizens, users and customers in times of crisis. This paper aims to analyse some of the different methods and scenario which can be applied to AI and big data, as well as the opportunities provided by the application in various business operations and crisis management domains.
Artificial Intelligence, Big Data, Business Operations, Crisis Management.
Atefeh Momenikorbekandi and Maysam F. Abbod, Department of Electronic and Electrical Engineering, Brunel University London, Uxbridge, UK
This paper applies an integrated genetic algorithm (GA) based on parthenogenetic algorithm to optimise job shop scheduling problems for single machine and multi-machine job-shops. The integrated GA has been tested with and without three parthenogenetic operators, namely swap, reverse and insert. The makespan of job shop scheduling problems refers to the total length of the schedule when all the jobs have been finished. The proposed GA algorithm utilises a combination of different types of GA selection functions, namely stochastic, roulette, sexual, and ageing, with parthenogenetic procedure which employs gene recombination instead of the traditional crossover operator in order to produce off springs. The proposed GA has been applied to single machine job shop and multi-machine with tardiness, earliness, and due date penalties.
Genetic Algorithm, single machine job shop, multi-machine job shop, parthenogenetic algorithm.
Pradhi Anil Kumar Das, Deepak Singh Tomar, Department of Computer Science & Engineering, Maulana Azad National Institute of Technology Bhopal, India
Lately, one of the most common illegal activities include the use of shooting weapons. In such dangerous situations, there is a dire need of preventive measures that can automatically detect such munitions. This paper presents the use of computer vision and deep learning to detect weapons like guns, revolvers and pistols. Convolutional Neural Networks can be efficiently used for object detection. In this paper, precisely, two Convolutional Neural Network (CNN) architectures - Faster R-CNN with VGG16 and YOLOv3, have been used, to carry out the detection of such weapons. The pre-trained neural networks were fed with images of guns from the Internet Movie Firearms Database (IMFDB) which is a benchmark gun database. For negative case images, MS COCO dataset was used. The goal of this paper is to present and compare performance of the two models to bring about gun detection in any given scenario. The results of YOLOv3 outperforms Faster R-CNN with VGG16. The ultimate aim of this paper is to detect guns in an image accurately which in turn can aid crime investigation.
weapon detection, gun detection, computer vision, deep learning, artificial intelligence.
Ruolin Huang, Ting Lu1, Yiyang Luo, Guohua Liu and Shan Chang, College of Computer Science and Technology, Donghua University, Shanghai, China 201620
Federated Learning (FL) is a setting that allows clients to train a joint global model collaboratively while keeping data locally. Due to FL has advantages of data confidential and distributed computing, interest in this area has increased. In this paper, we designed a new FL algorithm named FedRAD. Random communication and dynamic aggregation methods are proposed for FedRAD. Random communication method enables FL system use the combination of fixed communication intervals and constrained variable intervals in a single task. Dynamic aggregation method reforms aggregation weights and makes weights update automately. Both methods aim to improve model performance. We evaluated two proposed methods respectively, and compared FedRAD with three algorithms on three hyperparameters. Results at CIFAR-10 demonstrate that each method has good performance, and FedRAD can achieve higher classification accuracy than state-of-the-art FL algorithms.
federated learning, random communication, dynamic aggregation, self-learning.
Jivitesh Sharma, Center for Artificial Intelligence Research, University of Agder, Norway
All of the fault detection methods proposed so far have been very complex and complicated procedures, which makes the fault detection models very slow and rigid. Also, the classification problems formulated for fault detection have not been exhaustive enough. All of the above problems, i.e. simplicity, speed, accuracy and class distribution for fault detection have been addressed in this paper and a new methodology is proposed. The proposed methodology uses the simplest feature extraction technique, which is the statistical feature extraction and uses only one feature for feature extraction, instead of using 10-15 of them, which has generally been the norm in previous works. So, there is a huge reduction in the number of feature. The Extre me Learning Machine which is one of the fastest and most accurate learning algorithms is used for classification. Also, the classification problem for fault detection considered in this paper is the most exhaustive, consisting of 19 classes and distinguishes between fault type, fault size and bearing position (DE or FE). The results show that the proposed method achieves 99.6% accuracy with ELM using only one feature and also 5x to 6700x speed up compared to other algorithms like SVM and its variants, Decision Trees, KNN and its variants and Ensemble methods.
Fault Detection, Ball Bearing, Statistical Features, Extreme Learning Machine.
Mohammad Saleem and BenceKovari, Department of Automation and Applied Informatics, Budapest University of Technology and Economics, Budapest, Hungary
Online signatures are one of the most commonly used biometrics. Several verification systems and public databases were presented in this field. This paper presents a combination of k-nearest neighbour and dynamic time warping algorithm as a verification system using the newly published DeepSignDB database. It was applied on both finger and stylus input signatures which represents both once and mobile scenarios. The system first tested on the development set of the database and achieved 6.04% for the stylus input signatures, 5.20% for the finger input signatures, and 6.00% for a combination of both types. These results ranked second and first in the SVC2021 signature verification competition development phase for both scenarios. The system also achieved a bronze medal in the evaluation phase of the competition for the finger input signatures task.
Online signature verification, k-nearest neighbour, dynamic time warping.
Gereltsetseg Altangerel and Tejfel Máté, Department of Programming Languages and Compilers, ELTE, Eötvös Loránd University, Budapest, Hungary
By programming both the data plane and the control plane, network operators can customize their networks based on their needs, regardless of the hardware manufacturer. Control plane programming, a major component of the SDN (Software Defined Network) concept, has been developed for more than 10 years and successfully implemented in real networks. Efforts to develop reconfigurable data planes and highlevel network programming languages make it truly possible to program data planes. Therefore, the programmable data planes and SDNs offer great flexibility in network customization, allowing many innovations to be introduced on the network. The general focus of research on the data plane is data-plane abstractions, languages and compilers, data plane algorithms, and applications. This article outlines some emerging applications on the data plane and offers opportunities for further improvement and optimization.
Data plane applications, load balancing, in-network caching, in-network computing, in-network data aggregation, in-band network telemetry (INT).
Ms. Agampreet Kaur Walia1, Mr. Dakshraj Sharma2 and Dr. Amit Chhabra3, 1Computer Science, Chandigarh College of Engineering and Technology, India, 2Computer Science, Chandigarh College of Engineering and Technology, India, 3Faculty of Engineering, Chandigarh College of Engineering and Technology, India
Network simulations are a popular methodology for the testing and study of the behavior of Network systems under specific situations and with specific inputs without having to put the actual corresponding system at risk. These network simulations, carried out through network simulation software, tend to be much more realistic and accurate in their outcomes compared to basic mathematical or analytical models. In addition, the advantages of using network simulation for testing, feature planning, etc. over testing on the actual system are too many to count — savings in money and time, risk avoidance in case of failures, etc. Today, there are several competitive as well as feature-rich network simulators available in the market for researchers, each with its own strengths and merits. When deciding on a simulator to utilize to satisfy their purpose, researches may be led to compare the various positives and negatives of one network simulator over another. Our work intends to aid researchers in such a comparison, by providing an overview of the various features of some of the most popular network simulators available for use in the research community.
QualNet, PSIM, GrooveNet, PeerSim, Mininet.
Yipkei Kwok and David L. Sullivan, Department of Computer Science, Mathematics and Physics, Missouri Western State University, Saint Joseph, Missouri, USA
Recent machine learning-based caching algorithm have shown promise. Among them, Learning-From-OPT (LFO) is the state-of-the-art supervised learning caching algorithm. LFO has a parameter named Window Size, which defines how often the algorithm generates a new machine-learning model. While using a small window size allows the algorithm to be more adaptive to changes in request behaviors, through experimenting with LFO, we realized that LFOs performance suffers dramatically with small window sizes. In this paper, we proposed LFO2, an improved LFO algorithm, which achieves high object hit ratios (OHR) with small window sizes. Our results show a 9% OHR increase with LFO2.
Content Delivery Network, Cache, Machine Learning, Supervised learning.
Lijun Dong, Richard Li, Futurewei Technologies Inc., 2220 Central Expressway, Santa Clara, CA, USA
The services provided by the today’s Internet cannot satisfy the precise latency guarantee required by many emerging applications. With New IP framework, the paper proposes that the order of the packets in the outgoing queue could be deliberately manipulated in order to satisfy the deadline constraints for as many packets as possible while achieving the minimum average stay time in a router. The algorithms based on backtracking, branch and bound are proposed to address the optimal scheduling problem for multiple concurrent packets. The performance evaluation verifies that the proposed scheduling schemes can achieve the best performance on the in-time delivery of packets from multiple flows.
in-time guarantee, multiple packets, New IP, best effort, contract, metadata, high precision communication, QoS, precise latency, backtracking, branch and bound.
Rahil Gandotra1 and Levi Perigo2, 1Interdisciplinary Telecom Program, University of Colorado Boulder, USA, 2Department of Computer Science, University of Colorado Boulder, USA
The energy consumption of network infrastructures is increasing; therefore, research efforts designed to diminish this growing carbon footprint are necessary. Building on our prior work, which determined a difference in the energy consumption of network hardware based on their forwarding configurations and developed a real-time network energy monitoring tool, this research proposes a novel technique to incorporate individual device energy efficiency into network routing decisions. A new routing metric and algorithm are presented to select the lowest-power, least-congested paths between destinations, known as Green Power Forwarding (GPF). In addition, a network dial is developed to enhance GPF by allowing network administrators to tune the network to optimally operate between energy savings and network performance. To ensure the scope of this research for industry adoption, implementation details for different generations of networking infrastructure (past, present, and future) are also discussed. The experiment results indicate that significant energy and, in turn, cost savings can be achieved by employing the proposed GPF technique without a reduction in network performance.
Energy efficiency, intent-based networking, network optimization dial, SDN, programmable control plane, OpenFlow, programmable data plane, P4.
Susith Hemathilaka and Achala Aponso, Department of Computer Science, Informatics Institute of Technology, Colombo, Sri Lanka
The face mask is an essential sanitaryware in daily lives growing during the pandemic period and is a big threat to current face recognition systems. The masks destroy a lot of details in a large area of face and it makes it difficult to recognize them even for humans. The evaluation report shows the difficulty well when recognizing masked faces. Rapid development and breakthrough of deep learning in the recent past have witnessed most promising results from face recognition algorithms. But they fail to perform far from satisfactory levels in the unconstrained environment during the challenges such as varying lighting conditions, low resolution, facial expressions, pose variation and occlusions. Facial occlusions are considered one of the most intractable problems. Especially when the occlusion occupies a large region of the face because it destroys lots of official features.
CNN, Deep Learning, Face Recognition, Multi-Branch ConvNets.
Gilles Cohen1, Pascal Briot2 and Pierre Chopard2, 1Finance Directorate Geneva University Hospitals Geneva, Switzerland, 2Quality and Patient Safety Division Geneva University Hospitals Geneva, Switzerland
In hospitalized populations, there is significant heterogeneity in patient characteristics, disease severity, and treatment responses, which generally translates into very different related outcomes and costs. A better understanding of this heterogeneity can lead to better management, more effective and efficient treatment by personalizing care to better meet patient’s profiles. Thus, identifying distinct clinical profiles among patients can lead to more homogenous subgroups of patients. Super-utilizers (SUs) are such a group, who contribute a substantial proportion of health care costs and utilize a disproportionately high share of health care resources. This study uses cost, utilization metrics and clinical information to segment the population of patients (N=32,759) admitted to the University Hospital of Geneva in 2019 and thus identify the characteristics of its SUs group using Latent Class Analysis. This study demonstrates how cluster analysis might be useful to hospitals for identifying super-utilizers within their patient population and determining their characteristics.
Latent Class Analysis, Clustering, Super-Utilizers, Inpatient Segmentation, Hospital Efficiency, Quality Improvement.
Manav Agarwal, Shreya Venugopal, Rishab Kashyap, Department of CSE, PES University, Bangalore, India
The film industry is one of the most popular entertainment industries and one of the biggest markets for business. Among the contributing factors to this would be the success of a movie in terms of its popularity as well as its box office performance. Hence, we create a comprehensive comparison between the various machine learning models to predict the rate of success of a movie. The effectiveness of these models along with their statistical significance is studied to conclude which of these models is the best predictor. Some insights regarding factors that affect the success of the movies are also found. The models studied include some Regression models, Machine Learning models, a Time Series model and a Neural Network with the Neural Network being the best performing model with an accuracy of about 86%. Additionally, as part of the testing data for the movies released in 2020 are analysed.
Machine Learning models, Time Series, Movie Success, Neural Network, Statistical significance.
Paul Groß and Philipp Kappus, Department of Computer Engineering, Baden-Wuerttemberg Cooperative State University, Friedrichshafen, Germany
In this paper we present two clustering methods to determine users with similar opinions on the Covid-19 pandemic and the related public debate in Germany. We believe, they can help gaining an overview over similar-minded groups and could support the prevention of fake-news distribution. The first method uses a new approach to create a network based on retweet-relationships between users and the most retweeted characters (influencers). The second method extracts hashtags from users posts to create a “user feature vector” which is then clustered using a similarity matrix based on [1] to identify groups using the same language. With both approaches it was possible to identify clusters that seem to fit groups of different public opinion in Germany. However, we also found that clusters from one approach cannot be associated with clusters from the other due to filtering steps in the two methods.
Data Analysis, Twitter, Covid-19, Retweet network, Hashtags.
Christoph Kammerer1, Michael Gaust2, Pascal Starke2, Roman Radtke1and Alexander Jesser1, 1University of Applied Sciences Heilbronn, Max-Planck-Str. 39, 74081 Heilbronn, Germany, 2CeraCon GmbH, Talstraße 2, 97990 Weikersheim, Germany
Reducing costs is an important part in todays business. Therefore manufacturers try to reduce unnecessary work processes and storage costs. Machine maintenance is a big, complex, regular process. In addition, the spare parts required for this must be kept in stock until a machine fails. In order to avoid a production breakdown in the event of an unexpected failure, more and more manufacturers rely on predictive maintenance for their machines. This enables more precise planning of necessary maintenance and repair work, as well as a precise ordering of the spare parts required for this. A large amount of past as well as current information is required to create such a predictive forecast about machines. With the classification of motors based on vibration, this paper deals with the implementation of predictive maintenance for thermal systems. There is an overview of suitable sensors and data processing methods, as well as various classification algorithms. In the end, the best sensor-algorithm combinations are shown.
Predictive Maintenance, Industry 4.0, Internet of Things, Big Data, Industrial Internet, ARMA.
Bayan Aldashnan and Maram Alkhayyal, Information Systems Department, King Saud University, Riyadh, Saudi Arabia
Due to the need for extensively customized services and the rapid diffusion of technological innovations in artificial intelligence and deep learning, chatbots are a notable development that have been adopted across diverse sectors including e-commerce. A chatbot can provide customers with a more convenient way of receiving answers in a timely manner instead of waiting for a call center or e-mail while also utilizing customer care representatives in more critical tasks that require human involvement. Chatbots can be equipped with deep learning and natural language processing tools to automate customer service and facilitate communication which consequently enhances customer satisfaction and overall profits. This review provides a classification and assessment of recent state-of-the-art deep learning based chatbot systems for customer service.
Deep Learning, Chatbot, Customer Service.
Sunil Gundapu and Radhika Mamidi, Language Technologies Research Centre, KCIS, IIIT Hyderabad, Telangana, India
The exponential rise of social media networks has allowed the production, distribution, and consumption of data at a phenomenal rate. Moreover, the social media revolution has brought a unique phenomenon to social media platforms called Internet memes. Internet memes are one of the most popular contents used on social media, and they can be in the form of images with a witty, catchy, or satirical text description. In this paper, we are dealing with propaganda that is often seen in Internet memes. Propaganda is communication, which frequently includes psychological and rhetorical techniques to manipulate or influence an audience to act or respond as the propagandist wants. To detect propaganda in Internet memes, we propose a multimodal deep learning fusion system. Our approach fusing the text and image feature representations and outperforms individual models based solely on either text or images.
Social Media, Internet Memes, Propaganda, Multimodal Fusion, Language & Vision.
Dulitha Lansakara, Chamara Niroshana, Thinusha Gunasekera, Imali Weerasinghe, Department of Information Technology, Sri Lanka Institute of Information Technology, Malabe 10115, Sri Lanka
Sri Lanka has been witnessing an increase in kidney disease issues for a while. Elderly kidney patients, kidney transplant patients who passed the risk level after the surgery are not treated in the emergency clinic. These patients are handed over to their families to take care of them. In any case, it is impossible to tackle a portion of the issues that emerge regarding the patient at home. It is hoped to enter patient’s data from home every day and to develop a system that can use that entered data to predict whether a patient is in an essential circumstance or not. Additionally, individuals in high-hazard regions are unable to know whether they are in danger of creating kidney disappointments or not and individuals in danger of creating kidney sickness because of Diabetes Mellitus. Thus, we desire to emphasize the framework to improve answers for this issue. The research focuses on developing a system that includes early kidney disease prediction models involving machine learning classification algorithms by considering the relevant variables. In predictive analysis, six machine learning methods are used: Support Vector Machine (SVM with kernels), Random Forest (RF), Decision Tree, Logistic Regression, and Multilayer Perceptron. These classification algorithms performance is evaluated using statistical measures such as sensitivity (recall), precision, accuracy, and F-score. In categorizing, accuracy determines which examples are accurate. The experimental results reveal that Support Vector Machine outperforms other classification algorithms in terms of accuracy.
Chronic Kidney Disease, Support Vector Machine, Diabetes Mellitus, Random Forest.
Sylvain Muller and Ciaran Bryce, University of Applied Sciences (HES-SO), Geneva, Switzerland
Regular data backups are fundamental for protection against cyber-attacks and damage to infrastructure. However, the real challenge is not a successful backup, but rather to ensure a successful restoration. It is important to ensure that backed up data remains usable for restoration in the companys current environment. The paper proposes an automated test framework that validates the continued usability of backed up data for target restoration environments. The framework tests backups of Excel files, MySQL and Postgres databases, PDP documents and flat files.
Security, backup, automation, testing, infrastructure-as-code.
Abdelsalam M. Maatuk, Sohil F. Alshareef and Tawfig M. Abdelaziz, Faculty of Information Technology, University of Benghazi, Libya & Faculty of Information Technology, Libyan International Medical University, Libya
Requirement engineering is a discipline of software engineering that is concerned with the identification and handling of user and system requirements. Aspect-Oriented Requirements Engineering (AORE) extends the existing requirements engineering approaches to cope with the issue of tangling and scattering resulted from crosscutting concerns. Crosscutting concerns are considered as potential aspects and can lead to the phenomena “tyranny of the dominant decomposition”. Requirements-level aspects are responsible for producing scattered and tangled descriptions of requirements in the requirements document. Validation of requirements artifacts is an essential task in software development. This task ensures that requirements are correct and valid in terms of completeness and consistency, hence, reducing the development cost, maintenance and establish an approximately correct estimate of effort and completion time of the project. In this paper, we present a validation framework to validate the aspectual requirements and the crosscutting relationship of concerns that are resulted from the requirements engineering phase. The proposed framework comprises a high-level and low-level validation to implement on software requirements specification (SRS). The high-level validation validates the concerns with stakeholders, whereas the low-level validation validates the aspectual requirement by requirements engineers and analysts using a checklist. The approach has been evaluated using an experimental study on two AORE approaches. The approaches are viewpoint-based called AORE with ArCaDe and lexical analysis based on Theme/Doc approach. The results obtained from the study demonstrate that the proposed framework is an effective validation model for AORE artifacts.
AORE, Validation and Verification, Requirements Engineering, Aspectual Requirements, Crosscutting Concerns.
Pau Julià, David Salvador, Marc Peña, Department of Security, Typeform, Barcelona, Spain
Software development methodologies have evolved during the last years with the goal of reducing the time to market to the minimum possible. Agile is one of the most common and used methodologies for rapid application development. As the agile manifesto defines in its 12 principles, one of its main goals is to satisfy the customer needs through early and continuous delivery of valuable software. It is significant that none of the principles refers to security. In this paper, we will explain how Typeform integrates security activities into the whole development process, reducing at the same time the phases on the SSDLC to reduce friction and improve delivery maintaining the security level
Security, S-SDLC, SDLC, AGILE, Development, Methodology.
Ismail Olaniyi MURAINA and Imran Ademola ADELEKE, Department of Computer Science, Adeniran Ogunsanya College of Education,Otto/Ijanikin, Lagos, Nigeria
One of the pillars of research writing is research design. Students’ understanding of research design will enable such student(s) to know the kind of variable he/she has in hand and also be able to select appropriate instrument to collect the relevant data needed to complete the research. Once the design of the project is in vague to the researcher or graduating students writing research then the problem begins to set in. Research design gives the researcher the direction of the study in terms of research questions or hypotheses formulation, use of the instrument and the analysis type. This paper presents Research Design Selection – Expert (RDS-Expert) via WIN-PROLOG 6.000 and LPA Toolkits,which available to be employed in selecting appropriate research design for a desired study. The Visirule software was used as a decision supporting tool, using Logic Programming Model to present RDS-Expert in a concisely and precisely way. The RDS-Expert serves as a guide for researchers to use in making good decision regarding the type of research design fit their studies.
Research Design, Visirule, Logic Programming Model, Project writing.
Prof. Hiral M. Patel, Prof. Rupal R. Patel and Ms. Akshita A. Patel, Department of Computer Engineering, Sankalchand Patel College of Engineering, Sankalchand Patel University, Gujarat, India
Internet of Things (IoT) innovation can possibly be used to battle the COVID-19 pandemic. Shut spaces that are important for public or private workspaces like gathering rooms, lounge areas, homerooms, latrines, and so on are imparted to others in associations. These common spaces are obvious objectives for the spread of an infection and thus it turns out to be considerably more basic to know whether individuals who enter the workspace are debilitated. At the point when the entire world is battling a similar foe, we have all needed to accept new advancements and find their advantages, in spite of the fact that there is as yet far to go. Therefore, COVID-19 may well have been the epitome of the Internet of Things (IoT). Taking into account the current circumstance the COVID-19 has gotten pervasive in each edge of the world. We should target forestalling the local area spread of the infection. To accomplish this we should ensure a legitimate social distance is kept up from one individual to another. To ensure that an appropriate social distance is kept up from one individual to another we are concocting the possibility of social distancing shoes.
COVID-19, Coronavirus, Internet of Things, Social Distance, Pandemic.
Magesh Kasthuri, Hitarshi Buch, Krishna Moorthy, Vinod Panicker, Wipro Limited
Data is a new currency today and it is vulnerable to threat attack and data security is utmost important for any organization. The economics of data being used across the industries rapidly growing in current digital world so the potential data related threats is also rapidly growing. Data security is an integrated solution component for any Enterprise solution but with the growing demand on data security and potential threat handling, Data Security as a Service a.k.a DSaaS is a new model widely accepted in modern age architecture in Blockchain and Big Data world combining the power of cloud based security services, decentralized network in Blockchain and tamper-proof ledger management. Any Enterprise Security architecture comprises of how data is handled in a secured way and how integration between services (consumers/producers or API interaction or any middleware services) handles data between them. Hence it is inevitable to that future technology adoption should include Data Security-as-a-service for zero-trust solution design complying with compliance and security standards for industry.
Data Security, Blockchain, Decentralized Ledger, DSaaS, DLP, UEBA, CASB, Certificate Management, Key Management.
Mamoipone Elisa Masupa, Central University of Technology Free State, South Africa
The Fourth Industrial Revolution (4IR) is a new phenomenon that will impact human society drastically. It is complex, highly dynamic and constantly evolving at an ever-increasing pace. To date most of the research on the topic of the 4IR is focused on technological and scientific topics, with little to no work done on SMMEs resilience. The issue of the ability to survive and successfully compete in a turbulent business environment (i.e., resilience) is becoming more and more noteworthy within entrepreneurial, managerial and strategic studies. SMMEs constitute the backbone of the national economy as they create job opportunities, steer Gross Domestic Product (GDP) growth, encourage entrepreneurial innovation and contribute to increased exports of the nation.However, SMEs face a number of challenges that impede their ability to fully provide the much-needed boost to the economic development of countries. The lack of technology adoption is often cited as one of the key challenges of most SMEs in South Africa. As we stand at the cusp of the 4IR digital technologies dimensions such as Blockchain, big data analytics and cloud computing and many more have become the means and solutions to many of the SMMEs resilience. This study examines the effects of 4IR digital technologies dimensions such as (blockchain,big data analytics and cloud computing) on SMMEs resilience, mediation of strategies such as (information management strategy and change management strategy) and workflow process.
Fourth industrial revolution (4IR), Blockchain, Big data analytics, Cloud computing, SMMEs resilience.
Gao Xiao, Department of Software Engineering, Xi an University of Posts and Telecommunications, China
With the development of omics technology, more and more data will be generated in cancer research. Machine learning methods have become the main method of analyzing these data. Omics data have the characteristics of a large number of features and small samples, but features are redundant to some extent for analysis. We can use the feature selection method to remove these redundant features. In this paper, we compare two SVM-based feature selection methods to complete the task of feature selection. We evaluate the performance of these two methods on three omics datasets, and the results showed that the SVM-RFE method performed better than the pure SVM method on these cancer datasets.
Cancer Classification, Feature Selection, Support Vector Machines, Recursive Feature Elimination.
Saima Ishaq, Reshma Mahjabeen, Ruheena Mahveen, Department of Computer Science, King Khalid University Saudi Arabia, Abha
In the present era of big data, the overwhelming number of choices on the web has raised the need of prioritization, filtering and effective delivery of relevant information for alleviating the issue of information overload, which has resulted in problems to numerous web users. Various information filtering (IF) systems have been developed for tackling the issue of information overload for particular tastes. Different techniques and models from varied research areas such as behavioral science, artificial intelligence and information retrieval etc. are used by these systems. This paper is focused on highlighting the main concept of web information filtering and clarifying the difference between IF systems and related ones. Moreover, the IF techniques employed in available research studies are discussed and their pros and cons are analyzed in depth. The limitations of the current IF systems are described and some amendments are suggested for future, to enhance the procedure of information filtering on web and making web experience more productive and less time-consuming for the users.
Information filtering, Web information, User profile, Content-based Filtering, Collaborative filtering, Web mining.
Fabio Alexandre M.H. Silva, Bruno A. Bonifacio, Fabio Oliveira Ferreira, Fabio Coelho Ramos, Marcos Aurelio Dias and Andre Ferreira Neto, Sidia Institute of Science & Technology, Manaus, Amazonas, Brazil
Although Distributed Software Development (DSD) has been a growing trend in the software industry, performing requirements management in such conditions implies overcoming new limitations resulting from geographic separation. SIDIA is a Research and Development (R&D) Institute, located in Brazil, responsible for producing improvements on the Android Platform for Samsung Products in all Latin America. As we work in collaboration stakeholders provided by Mobile Network Operators (MNO) from Latin countries, it is common that software requirements be provided by external stakeholders. As such, it is difficult to manage these requirements due to the coordination of many different stakeholders in a distributed setting. In order to minimize the risks, we developed a tool to assist our requirements management and development process. This experience paper explores the experience in designing and deploying a software approach that facilitates (I) Distributed Software Development, (II) minimizes requirements error rate, (III) teams and task allocations and (IV) requirements managements. We also report three lessons learned from adopting automated support in the DDS environment.
industrial case study, requirement management, DSD, distributed software development, RM, automation, industrial experience.
Thallys Rubens Moreira Costalat1 and Géssica Fortes Tavares2, 1Institute of Technology, Federal University of Pará, Pará, Brazil, 2Institute of Health Science, Federal University of Pará, Pará, Brazil
Currently, one of the leading causes of death around the world are caused by diseases or acute syndromes installed in the cardiovascular system of the human body. Thus, this paper presents a modern alternative for the detection of cardiovascular diseases from health indicators such as age, gender, glucose and cholesterol indices, used as inputs for machine learning systems. The evaluation is made by using supervised learning algorithms, such as K-Nearest Neighbours, Decision Tree, Logistic Regression, Voting Classification, from the accuracy observed during the testing period, in order to conclude what is the best alternative for the construction of an effective cardiovascular event predictor in the clinical routine.
Machine Learning, Health Science, Supervised Learning, Heart Disease, Scikit-Learn, Python.
Khadidja Henni1,2, Lina Abou-Abbas1,2, Imene Jmel1,3, Amar Mitiche3, Neila Mezghani1,2, 1Imaging and Orthopaedics Research Laboratory, The CHUM Research Center, Montreal, Canada, 2Research Institute LICEF, TELUQ University, Montreal, Canada, 3INRS-Centre Énergie, Matériaux et Télécommunications, Montréal, Canada
Automatic epileptic seizure detection is a challenging task that could cope with sudden seizures and help epileptic patients to have a normal life. The electroencephalography (EEG) recording remains the most common method used for detecting epileptic seizures. The precision and accuracy of seizure detection are the most important elements in automatic EEG-based seizure detection systems, which could be achieved by training the classification models with relevant features. In this work, we propose a robust machine learning framework for epileptic seizure detection from EEG data. Imbalance class problem and high dimensional feature space issue have been handled for classification. Our approach has been tested on the largest EEG database (The Temple University Hospital EEG Seizure Corpus, TUSZ). A comparative study on three categories of data balancing techniques: cost-sensitive learning (weighting), oversampling and under sampling has been made. An efficient feature selection algorithm based on feature interaction graph analysis has been used for selecting minimal number of relevant inputs before classification. Results in terms of accuracy and area under the curve (AUC), have showed that the features subset selected using the graph-based method, balanced by the Synthetic Minority Over Sampling method (SMOTE) achieved the highest classification performance using random forest classifier.
Epileptic seizure detection, EEG, Graph analysis, SMOTE, imbalanced classes, Random Forest.
Arwa Alrawais1, Fatemah Alharbi2, Moteeb Almoteri3, Sara A Aljwair4 and Sara S Aljwair5, 1,4,5College of Computer Engineering and Sciences, Prince Sattam Bin Abdulaziz University, Alkharj,16278, Saudi Arabia, 2College of Computer Science and Engineering, Taibah University, Yanbu 46522, Saudi Arabia, 3College of Business Administration, King Saud University, Riyadh, 11451, Saudi Arabia
The COVID-19 pandemic has swapped the world, causing enormous cases, which led to high mortality rates across the globe. Internet of Things (IoT) based social distancing techniques and many current and emerging technologies have contributed to the fight against the spread of pandemics and reduce the number of positive cases. These technologies generate massive data, which will pose a significant threat to data owners’ privacy by revealing their lifestyle and personal information since that data is stored and managed by a third party like a cloud. This paper provides a new privacy-preserving scheme based on anonymization using an improved slicing technique and implying distributed fog computing. Our implementation shows that the proposed approach ensures data privacy against a third party intending to violate it for any purpose. Furthermore, our results illustrate our scheme’s efficiency and effectiveness.
Anonymization, Fog computing, IoT, Privacy, Social distancing technologies.
Issam Moghrabi, Gulf University for Science and Technology, Kuwait
This paper explores the impacts of digital technologies on supply chains and coordination, the manufacturing process, energy conservation, efficiency, and environmental conservation. Digital transformation has led to the popularization of sustainable manufacturing, which entails creating sustainable products that promote social, economic, and environmental sustainability. Digital transformation has boosted sustainability in production and manufacturing in a variety of ways. These ways include increasing cross-border communication through the internet, decentralizing supply chains, Internet of Things (IoT) solutions, artificial intelligence, machine learning, big data analytics in predictive analysis, robotics, horizontal and vertical integration of businesses, efficient management, and various other ways. The findings of the paper indicate that digital transformation has changed manufacturing in various ways. Aspects like cloud computing, vertical and horizontal integration, communication, and the internet have contributed to sustainable manufacturing by decentralizing supply chains. In addition, some digital transformation tools such as predictive analysis and big data analytics have helped optimize sustainable manufacturing by reducing overproduction or underproduction through predicting customer demands.
Internet of Things, Digital Transformation, Machine Learning, sustainable organization.
Jordan Raykov, JDElite Consulting,Boulder, Colorado, USA
This paper presents a method for automated orthogonal edge routing of directed layered graphs usingthedescribededge crossings reduction heuristic algorithm. The method assumes the nodes are pre-arranged on a rectangular grid consisting of layers across the flow direction and lanes along the flow direction. Both the layers and lanes are separated by rectangular regions defined as pipes. Each pipe has associated segment tracks. The edges are routed along the shortest paths as orthogonal polylines composed of chained line segments. Each segment is assigned to a pipe and to a segment track in it. The edge crossings reduction phase uses an iterative algorithm to resolve crossings between segments. Conflicting segments are repositioned on adjacent segment tracks, either by swapping with adjacent segments, or by inserting additional tracks while considering the shortest paths of edges. The algorithm proved to be efficient and was implemented in an interactive graph design tool.
Directed Graphs, Orthogonal Edge Routing, Crossings Reduction Algorithms.
Igli Tafa, Fotion Konomi, Denis Koleci, Kledisa Gjuta, Polytechnic University of Tirana, Faculty of Information Technology, Information Engineering Department
During the last decade, different companies are thriving on re-search and development for creating full self-driving cars. To reach this state, the parts inside of a vehicle are communicating with each other harder than ever before and more securely than ever before. Therefore, the CAN Bus, which con-nects all inner electronic components inside of a car, has been developing at a rapid pace. During this development, one of the key aspects has been and will al-ways be security. CAN is a protocol that had no security mechanisms built in when it was invented. But as the components in the inner network of vehicles started to exchange more and more sensitive data, and furthermore, this network even started to communicate with outside networks, the CAN Bus was seriously threated by cyberattacks. In this paper we analyze many vulnerabilities of the CAN Bus and what countermeasures have been implemented to tackle every vulnerability, together with their advantages and their disadvantages. We will make a short recap for what is the CAN Bus, and which are the internal com-ponents that connect to it. Then, we will see some of the attacks that can be performed on the CAN Bus and then how some of the countermeasures, e.g., cryptography, intrusion detection systems deal with the attacks and their limita-tions.
CAN, Bus, ECU, Cyberattacks, Cryptography, Intrusion Detec-tion Systems, Sensors, Security.
Rahil Gandotra1 and Levi Perigo2, 1Interdisciplinary Telecom Program, University of Colorado Boulder, USA, 2Department of Computer Science, University of Colorado Boulder, USA
Energy consumption by the network infrastructure is growing expeditiously with the rise of the Internet. Critical research efforts have been pursued by academia, industry and governments to make networks, such as the Internet, operate more energy efficiently and reduce their power consumption. This work presents an in-depth survey of the approaches to reduce energy consumption in wired networks by first categorizing existing research into broad categories and then presenting the specific techniques, research challenges, and important conclusions. At a broad level, we present five categories of approaches for energy efficiency in wired networks – (i) sleeping of network elements, (ii) link rate adaptation, (iii) proxying, (iv) store and forward, and (v) network traffic aggregation. Additionally, this survey reviews work in energy modeling and measurement, energy-related standards and metrics, and enumerates discussion points for future work and motivations.
Energy efficiency, energy proportionality, energy-aware protocols, wired networks.
William da Rosa Frhlich, UNISINOS University, Brazil
Wearables sensors are essential devices that can get reliable physiological signals for disease diagnostics pattern identification. Few studies evaluate the effect of combining several different signs, pattern detection architectures, and the implication of wearable sensor data acquisition procedures. This paper aims to investigate the possible integration of data obtained from heart rate variability (HRV), electrocardiographic (ECG), electrodermal activity (EDA) electromyography (EMG), blood volume pulse (BVP), respiration changes (RSP), body temperature (BT) and three-axis acceleration (ACC) to detect stress patterns. We compared using different machine learning and deep learning architectures among different architectures and datasets. The developed model shows promising results. In order to test the proposed model, we used two different datasets. The results obtained during the model training varied between 60 % for four classes and 87% accuracy for two classes.
Wearables, Deep Learning, Physiological Signals, Diagnostics.
Thomas Burton, University of Oxford, United Kingdom
In this paper, we propose a new method for fingerprinting different classes of wireless devices. Our method relies on the observation that different device types, or indeed different models of the same type (e.g., different models of smartphones), have different wireless radiation patterns. We show in detail how a small set of stationary receivers can measure the radiation pattern of a transmitting device in a completely passive manner. As the observed device moves, our method can gather enough data to characterize the shape of the radiation pattern, which can be used to determine the type of the transmitting device from a database of patterns. We apply this idea to the problem of identifying the model of smartphones present in an office environment. We demonstrate that the patterns produced by different models of smartphones are easily different enough to be identified. Our measurements are repeatably measurable using RSS with commercial-off-the-shelf hardware.
Wireless Radiation Patterns, Device Fingerprinting, Identification.
