Advances in Intelligent Networking and Collaborative Systems

Research in the area of SaaS revealed that several different aspects are critical to the usage of such cloud services. The drawback is that usually all of the identified success factors seem to be equally important. Therefore, this paper provides insights from a survey on the ranking of 18 relevant success factors according to their perceived actual performance and priority among 274 individuals. The results show that the priority of several factors like the image of the adopter, a regulatory framework, observability, trialability, and energy efficiency are perceived as sub relevant. The three factors with the highest priority are security, trustfulness, and ease of use of the service.

To understand how proteins function it is crucial to understand the connection between their structure, flexibility and dynamics. In the field of bioinformatics and computational biology, there is a strong interest to develop software and computational tools that analyzes various properties of protein structures. The software we are using for protein flexibility and dynamics analysis generally assumes a single task, single thread environment. To more efficiently elucidate the function of proteins, we need to perform large-scale calculations on many structures. To improve computational speed of such large scale analysis, we decided to perform parallel distributed computation with the conventional software. We designed a simple protocol dedicated to this software over http and achieved a speedup of 550 times with 600 CPU cores. With such speed ups, we are able to perform faster high-throughput computations on large number of protein structures.

Many of us never start thinking about secure information communication and collaboration until it is too late, after a security incident has been identified. The problem is that awareness of the issue remains quite low, coupled with a lack of continuing education in the field. Research on the subject and the level of security awareness was undertaken at Comenius University in Bratislava by the Faculty of Management at the end of 2017 and beginning of 2018, with a wide range of organisations participating. A number of interesting results came out of the analysis, all of which need to be followed up. Our paper discusses the outcome of experiments related to the study’s findings, while pointing out possible solutions in the form of process changes and the use of cloud-based tools and technology in order to eliminate risks and potential threats.

The integration of Smart Grid (SG) with cloud and fog computing has improved the energy management system. The conversion of traditional grid system to SG with cloud environment results in enormous amount of data at the data centers. Rapid increase in the automated environment has increased the demand of cloud computing. Cloud computing provides services at the low cost and with better efficiency. Although problems still exists in cloud computing such as Response Time (RT), Processing Time (PT) and resource management. More users are being attracted towards cloud computing which is resulting in more energy consumption. Fog computing is emerged as an extension of cloud computing and have added more services to the cloud computing like security, latency and load traffic minimization. In this paper a Cuckoo Optimization Algorithm (COA) based load balancing technique is proposed for better management of resources. The COA is used to assign suitable tasks to Virtual Machines (VMs). The algorithm detects under and over utilized VMs and switch off the under-utilized VMs. This process turn down many VMs which puts a big impact on energy consumption. The simulation is done in Cloud Sim environment, it shows that proposed technique has better response time at low cost than other existing load balancing algorithms like Round Robin (RR) and Throttled.

In recent years, next generation power semiconductor devices using semiconductors with large band gap such as SiC (Silicon Carbide) attract attention. It is very important to detect crystal defects, surface processing defects including polishing, defects contained in the SiC substrate, defects included in the epitaxial growth film, defects caused by the device forming process, and so on. This is because elucidating the cause of the detected defect and investigating the influence on device quality and reliability lead to development of a better manufacturing method. Recently, observation with a low energy scanning electron microscope (LE-SEM) which is more accurate than C-DIC and PL has been put to practical use. As a result, crystal information of just below the outermost surface can also be obtained. However, since image processing techniques targeting SEM images of SiC substrates have not existed so far, it has not been possible to efficiently and automatically extract defects from enormous amounts of data. In this paper, we propose a method for detecting defects on SiC substrate by SEM and classifying them using deep learning.

Social Networks Analysis has become a common trend among scholars and researchers worldwide. A great number of companies, institutions and organisations are interested in social networks data mining. Information published on many social networks, like Facebook, Twitter or Instagram constitute an important asset in many application fields, overall sentiment analysis, but also economics analysis, politics analysis and so on. Social networks analysis comprehends many disciplines and involves the application of different methodologies and techniques to define the criteria for generating the analytics, according to the purpose of the study. In this work, we focused on the semantic analysis of the content of textual information obtained from social media, aiming at extracting hot topics from social networks. We considered, as case study, reviews from the Yelp social network. The same methodology can be also applied for social and political opinion mining campaigns.

In today’s dynamic environments, user feedback data are a valuable asset providing orientations about the achieved quality and possible improvements of various products or services. In this paper we will present a hybrid fuzzy clustering model combining variants of fuzzy c-means clustering and density based clustering for exploring well-structured user feedback data. Despite of the multitude of successful applications where these algorithms are applied separately, they also suffer drawbacks of various kinds. So, the FCM algorithm faces difficulties in detecting clusters of non-spherical shapes or densities and moreover it is sensitive to noise and outliers. On the other hand density-based clustering is not easily adaptable to generate fuzzy partitions. Our hybrid clustering model intertwines density-based clustering and variations of FCM intending to exploit the advantages of these two types of clustering approaches and diminishing their drawbacks. Finally we have assessed and compared our model in a real-world case study.

In this paper will be presented new classes of authentication procedures, based on visual CAPTCHA solutions, which require special cognitive skills or expert-knowledge. Such authentication protocols can be used especially to differentiate access possibilities for particular group of persons, based of theirs expertise. For new authentication protocols some possible examples of applications will be presented with relation to Cloud computing and distributed authentication.

The mandate of citizens for more socially responsible information systems that respect privacy and autonomy calls for a computational and storage decentralization. Crowd-sourced sensor networks monitor energy consumption and traffic jams. Distributed ledgers systems provide unprecedented opportunities to perform secure peer-to-peer transactions using blockchain. However, decentralized systems often show performance bottlenecks that undermine their broader adoption: propagating information in a network is costly and time-consuming. Optimization of cost-effectiveness with supervised machine learning is challenging. Training usually requires privacy-sensitive local data, for instance, adjusting the communication rate based on citizens’ mobility. This paper studies the following research question: How feasible is to train with privacy-preserving aggregate data and test on local data to improve cost-effectiveness of a decentralized system? Centralized machine learning optimization strategies are applied to DIAS, the Dynamic Intelligent Aggregation Service and they are compared to decentralized self-adaptive strategies that use local data instead. Experimental evaluation with a testing set of 2184 decentralized networks of 3000 nodes aggregating real-world Smart Grid data confirms the feasibility of a linear regression strategy to improve both estimation accuracy and communication cost, while the other optimization strategies show trade-offs.

Wireless Mesh Networks (WMNs) have many advantages such as low cost and increased high-speed wireless Internet connectivity, therefore WMNs are becoming an important networking infrastructure. In our previous work, we implemented a Particle Swarm Optimization (PSO) based simulation system for node placement in WMNs, called WMN-PSO. Also, we implemented a simulation system based on Simulated Annealing (SA) for solving node placement problem in WMNs, called WMN-SA. In this paper, we implement a hybrid simulation system based on PSO and SA, called WMN-PSOSA. We evaluate the performance of WMN-PSOSA by conducting computer simulations considering Normal and Uniform distributions of mesh clients. Simulation results show that WMN-PSOSA performs better for Normal distribution compared with the case of Uniform distribution.

Autonomous vehicles are slowly being modified to be cognitive. A fully autonomous vehicle is a robotic vehicle that is designed to travel to destinations without human intervention. It uses vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I), and vehicle-to-everything (V2X) communications in order to improve driving safety and traffic efficiency, and to provide information and entertainment to the driver. Its communication protocols are capable of driving without any human actions. V2X communications are designing and running in cellular networks. Some of the most commonly used methods for increasing the spectral efficiency of cellular systems involve resource allocation schemes. Various channel allocation schemes have been introduced to provide quality of service (QoS). To get high QoS, we have to examine the QoS requirements for vehicle cooperative communication systems. This paper discusses, analyzes, and evaluates the performance of a cooperative communication strategy for cellular systems, for applications in autonomous vehicle networks. The aim is to reduce delays in message exchange between vehicles that are caused by heavy traffic loads, a high blocking rate, and lengthy processing times. The major goal of our work is to design a robust message protocol for exchanging information among vehicles.

In recent years, some computer graphics (CG) animations in which musical instrument performance itself is important in the story, have attracted attention, and it is necessary to actually draw performance scenes by musical instruments. If we can automatically and semi-automatically generate CG animations including the performance of musical instruments, we can drastically reduce time and cost of production. For this reason, our research group has been advancing research and development of a system for generating CG animations of performance by musical instruments. In particular, the method of generating CG animations of performance by a plurality of musical instruments, such as a band performance and an orchestra performance, is necessary. For that purpose, we need to develop a method to detect onset times of each instrument from WAV format data separately recorded for each musical instrument and to synchronize the sound and the motion of all musical instruments. In this paper, we propose a method of generating a CG animation synchronizing motion and sound of multiple musical instruments.

Number of devices needing wireless communication is more than ever. The main ingredient of wireless communication-the spectrum-is limited and due to the unprecedented growth of smart devices the demand for spectrum is high. To serve all the devices needing wireless communication, intelligent spectrum sensing and its trading has been the hot topic of research in this decade. In spectrum trading, so far in the literature, two layers are considered in terms of primary and secondary users. However, it may be the case, that the secondary users (may be NGOs) may redistribute their spectrum to some third party (downtrodden people of the rural areas) freely. To the best of our knowledge, this environment is not addressed in the literature so far. In this paper this three layer potentially demanding architecture is studied and algorithms are proposed based on the theory of mechanism design without money. Our algorithm is also simulated with a specially designed benchmark algorithm.

This paper proposes collaborative laser beam alignment system of bilateral free space optics (FSO). Active free space optics apparatus is designed and prototyped which aligns transmission laser beam precisely. Adjustment statics is also investigated involving two techniques. A collaborative adjustment method makes it possible to connect the remote FSO apparatus using nothing but the transmission laser beam. Analytic estimation method of laser spot distribution is formulated based on the Gaussian beam optics. It helps the laser beam to reach the target receiver without feedback control.

The mathematical properties of the spatial movement patterns of animals, humans, and insects have gradually become clear in recent years. Motion tracking is essentially necessary for the study of the spatial movement patterns. GPS telemetry is often used for large mammals, birds, and humans. However, it is difficult to track the migration paths of insects and small fish by using GPS telemetry. When the region of an object’s movement is restricted, we can record the movement of the object by the video instead of GPS telemetry, but we must determine the position coordinates of objects from a video for the study of the spatial movement patterns. If this motion tracking can be executed not by manually but automatically, we can obtain and analyze Big data on motion. In this paper, we develop a system for motion tracking of one or more small fishes in an aquarium. This system solves the difficulties such as the overlap of fishes, the ghost image of the reflection, and outputs the trajectory of the 3-dimensional coordinates. Furthermore, we apply this system to actual videos and show that the detection of active and inactive phases is possible and that the spatial movement pattern follows Levy walk.

In the context of transportation technologies for smart city, recently, self-driving and connected cars have been studied with increased attention to realize more efficient and safer transportation. On the other hand, applications of blockchain and its Layer 2 technologies to IoT and cyber-physical systems have been considered for achieving sharing economy. The blockchain explores new incentive-based value exchange system using reliable data structure with resistance to information tampering despite the absence of reliable central control. In this paper, we propose a distributed system for collaborative traffic information gathering and sharing incentivized by blockchain technology. I consider that beacon devices are deployed along road segments by many and unspecified users. These devices collaboratively estimate traffic count and status, and also detect traffic jam and accidents to gather and share them coordinated by the incentive from its digital currency. After introducing an idea of how to realize this system, I demonstrate numerical simulations to evaluate some preliminary results for how the proposed system works and responds to real-time traffic conditions, such as traffic accident and jam in the road segment.

Information and communication technologies have important role in many different innovations and further positively affect the productivity and economic growth. Our aim was to examine examine the share of ICT sector in selected European countries and identify its potential determinants and consequences. We used mostly panel data for EU28 countries plus Norway and Switzerland and applied correlation and panel Granger causality tests. The employment in ICT was used as main examined variable. Our results suggest that there is a positive correlation between ICT skills as well as the share of people with ICT education and employment in ICT. The causal effect seems to be working in direction from education to employment. Moreover, variables capturing GDP per capita, trade openness, R&D expenditure, number of internet users, quality of regulation and political stability appear to be all positively correlated with employment in ICT and especially with employment in ICT services.

Along with the evolution of software development models, the role of tests in the whole process has been increasing, mainly in the direction of earlier detection of errors than at the time of final product delivery by the customer. The innovativeness and volatility of the Internet applications market leads to a situation where there is great pressure to quickly implement a good quality solution. It is crucial to ensure both high quality and effective monitoring, which will allow to decide on the implementation of the application at the right time. The paper concerns the problem of software testing and its automation, with particular emphasis on web applications created in Java.

Cloud is an on-demand centralized global internet service provider to the end-user. Cloud computing, however, faces problems like high latency and low degree of security and privacy. For low latency, better control of the system and high-security fogs are integrated in the architecture. The cloud-fog based architecture provides security, control of data, quick response and processing time. Recently one of the emerging research areas of Smart Grid (SG) is the integration of Internet-of-Things (IoTs) with SG services to improve its capability. IoTs are interrelated digital machines, objects, and computing devices which have the ability to transfer information over the internet without human interaction with the system. SG is a modern energy management grid for smart use of resources and to optimize Peak Average Ratio (PAR) of energy consumption. Cloud-fog based architecture is integrated with SG for efficient utilization of resources and better management of the system. In cloud-fog computing, the load balancer allocates requests of end-user to Virtual Machines (VMs). In this paper a load balancing scheduling algorithm is presented namely; Min-conflicts scheduling algorithm. The algorithm takes a heuristic approach to solve a Constraint Satisfaction Problem (CSP).

The use of the traditional grid with Information and Communication Technology (ICT) gave birth to Smart Grid (SG). New services and applications are built by utility companies to facilitate electricity consumers which generate a huge amount of data that is processed on the cloud. Fog computing is used on the edge of the cloud to reduce the load on cloud data centers. In this paper, a four-layered architecture is proposed to reduce electricity shortage between consumer and electricity providers. Clusters layer consist of clusters of buildings which are connected to Micro Grid (MG) layer. MG layer is further connected to fog and cloud layer. Three load balancing algorithms Round Robin (RR), Honey Bee Optimization (HBO) and State-Based Load Balancing (SBLB). Results demonstrate that SBLB outperforms RR and HBO in terms of Response Time (RT) and Processing Time (PT).

There are various methods and approaches to electronic voting all around the world. Each is connected with different benefits and issues. One of the most important and prevalent problems is lack of auditing capabilities and system verification methods. Blockchain technology, which recently gained a lot of attention, can provide a solution to this issue. This paper presents Auditable Blockchain Voting System (ABVS), which describes e-voting processes and components of a supervised internet voting system that is audit and verification capable. ABVS achieves this through utilization of blockchain technology and voter-verified paper audit trail.

The perception that every person has about her health is very important for health, health evolution and therefore for enjoying a healthy life. So, it is important to know what the factors that most influence self-rated health are. If these factors are known, changes that have long term impact on people’s health can be proposed, even for unhealthy people. The present paper performs a preliminary study in that direction, analyzing the causes of self-rated health from a couple of datasets: one private with 4848 people and one public with 802 people. These datasets have been analyzed to find out what socioeconomic, biological and environmental factors have more influence on health status and a dashboard has been created to allow analyzing the data interactively. Results show some factors that influence self-rated health such as chronic diseases, limitation in daily activities and depression were important.

The eHealth represents a system, collecting data related to the evidence of the health care. Ensures the interoperability of patient’s health history, medication and medical records, related to the care providers. Due precise evidence and data accessibility, this system definitely supports the fraud blocking, costs cutting and contributes to improvement of condition for better care provisioning by care providers, EU countries do not provide public data, targeted to clarify the results of eHealth in form of structured data set, representing the cost savings. This paper represents an alternative view to estimate the results of eHealth implementation into large measures health care system.

Maturity models have been introduced, over the last four decades, as guides and references for Information System (IS) management in organizations from different sectors of activity. In the educational field, maturity models have also been used to deal with the enormous complexity and demand of Educational Information Systems. This article presents a research project that aims to develop a new comprehensive maturity model for the Higher Education Institutions (HEI) area. The HEIMM (Higher Education Institutions Maturity Model) will be developed to help HEI to address the complexity of there IS, as a useful tool for the demanding role of the management of these systems, and institutions as well. The HEIMM will have the peculiarity of congregating a set of key maturity influence factors and respective characteristics, enabling not only the assessment of the global maturity of the HEI IS but also the individual maturity of its different dimensions. In this article, we present the second phase of our project by discussing the methodology for the development of maturity models that will be adopted for the design of the HEIMM and the underlying reasons for its choice.

Networks and information systems play a crucial role in free movement and are often interconnected and connected to the Internet as a global tool. Their disturbance in one Member State therefore affects other Member States and the whole European Union. Resilience of networks and stability of the information system is therefore a basic prerequisite for a smooth and undistorted functioning of the European Union’s internal market and a prerequisite for international cooperation. The National Security Authority, as the central authority for cyber security, has prepared a bill on Cyber Security on the basis of several approved documents. By its approval and effect from 1 March 2018, the content of Directive 2016/1148 of the European Parliament and of the Council from 6 July 2016 measures to ensure a high common level of network security and information systems in the Union and is also transposed into the law of the Slovak Republic. The aim of this paper is to investigate selected legal institutes of the law using a number of scientific methods and to point out their crucial importance for the cyber security of the Slovak Republic. The subject matter of the survey is the field of law of information and communication technology, military and administrative law. However, this is a multidisciplinary examination of this problem, which interferes with security management.

Hardly a week goes by without headlines about new cyber-attacks. As the sophistication of cyber-attacks constantly increases, organizations have to consider to be affected by attacks. In order to effectively and efficiently react to an incident, professional and well-organized incident management has to be in place. The major goal of this paper is to support organizations to develop and improve their cyber-security incident management. Therefore, in this work, a readiness model, covering nearly 80 topics and 500 requirements in the domain of incident management, is introduced.

It is well known that many actual networks have the scale-free property that the degree distribution follows the power law. This property is found in many actual networks in the real world such as the Internet, WWW, a food chain network, an airline network, and a human relations network. As for a generation mechanism of a human relations network, we should consider the human mobility that, in general, a person moves around, meets another person, and makes friend relations stochastically. However, there are few models considering human mobility so far. In this paper, we propose a new model that generates a human relations network considering human mobility, and we show that this model has the scale-free property by numerical experiments.

Crowdfunding has gained increased attention as an alternative type of investment opportunity. However, the rationale and the drivers of why this form of funding has grown quickly is partially unexplored. Engaging the crowd and capturing the wisdom of many is a growing phenomenon, which benefits entrepreneurs opening an alternative funding route as well as a path to knowledge acquisition. The entrepreneurial perspective when selecting a crowdfunding platform and the resulting decision parameters is the main topic of this research. By using the method of semi-structured interviews the research (i) analyzes the question what makes the entrepreneur trust in the online platform, and (ii) provides a deeper understanding of which kind of risk the entrepreneur has to face.The findings suggest, that both experience factors such as design, relevant content, easiness of navigation, security and customer feedback, as well as sensitivity factors including reliability, intention, integrity, competence, openness and perceived concern are major trust drivers which lead to a higher use of crowdfunding web platform.The components of perceived risk – in particular the risk of losing intellectual property (IP), the threat of a cyber-attack, and the risk of loss of autonomy by being dependent on a third party – were further observed results of analyzing the entrepreneurial perspective.

It is important to design robust networks that are resistant to network failures, because high reliability is required for communication networks as important infrastructure. For that purpose, there is an approach of link protection to decrease link failure probability by the backup mechanism and the fast recovery mechanism; however, enormous cost is required, if the failure probability of all links must be decreased. It is realistic to preferentially protect only some highly required links so that the reliability of the entire network is improved. In this paper, we assume that the failure probability of each link can be reduced according to cost for protection. Since the failure probability of each link is decreased by the cost allocated to the link, the network reliability defined as the probability that the entire network is connected, is increased. We define the network design problem that determines the cost allocated for each link, under the constraint that the sum of the cost allocated for each link must be less than or equal to a given threshold, so that the network reliability is maximized. We show the NP-hardness of this optimization problem and design a polynomial-time heuristic algorithm. Moreover, we evaluate the performance by using the topology of some actual communication networks, and we show that the proposed algorithm works well.

The primary aim of food traceability is to increase food safety but traceability systems can also bring other benefits to production systems and supply chains. Technology has helped supply chain distribution to be successful and more competitive in the global market by improving their customers satisfaction level and providing more reasonable prices compared to their competitors. However, nowadays supply chains are long and complex and include many different actors, beginning with the farmers, followed by collectors, traders, manufacturers. At the end, the processed products are difficult to trace back to their origins in a trusted way. Best way so far is the continuous medial or legal documentation of the chain of transactions, available only to a limited circle of persons. The recent efforts toward blockchain-based trust management open a new perspective here. Here, we propose a solution to link legal documentation and blockchain technology within a traceability system, with the specific application case of cacao and chocolate production in mind. For improved data security, the common blockchain concept is extended to a two-factor blockchain, where both blockchains are connected via digital watermarking of the documentation media, one blockchain traces the documentation steps, the other the watermarking embedding. By a unique reading-verification approach, the integrity of their relatedness can be proven, thus improving also reliability of the whole product tracing.

Transmission Control Protocol (TCP) with a loss-based congestion control is still dominantly used for reliable end-to-end data transfer over diverse types of network although it is ineffective when traversing lossy networks. We previously proposed an IP tunneling system across lossy networks using the TCP with Network Coding (TCP/NC tunnel) and showed its potential to significantly mitigate the goodput degradation of end-to-end TCP sessions without any change of end-device’s communications protocol stack, but it was shown only in homogeneous conditions. On the other hand, reliable end-to-end data transfer in diverse and heterogeneous IoT environments in a cost-efficient manner is an emerging challenge. Therefore, in this paper, we investigate the characteristics of the TCP/NC tunnel on heterogeneous networks with/without network congestions, to assess the applicability of the TCP/NC tunnel-based intelligent gateway system to IoT environments where end-devices are connected to a gateway with different link bandwidths or connected to different gateways in terms of network topology. The simulation results suggest the TCP/NC tunnel can efficiently utilize the bottleneck bandwidth in such heterogeneous situations even with congestion and achieve a significantly high goodput of end-to-end TCP sessions in a wide range of link loss degree especially when the tunnel link bandwidth is sufficient.

It is difficult to completely prevent failures of links and nodes in a communication network. Therefore, it is necessary to quickly recover the failed nodes. Since resources are generally limited at the time of a disaster, it is impossible to recover all failure nodes at the same time, and it must be recovered sequentially. However, as is often seen in recent large-scale earthquakes, the aftershocks often cause failures, give large damage again and again, and prevent the recovery. In such a situation, since a new failure may occur at any time during sequentially recovering many failures, the order of recovery significantly affects the communication quality of the network. Therefore, it is important to determine an appropriate recovery order so that the communication quality of a network keeps even in a situation that failures occur successively. In this paper, we propose a method of determining the node recovery order that as many nodes as possible keep to be connected even if successive node failures occur. First, we formulate this problem as an optimization problem, prove that the problem is NP-hard, and design a heuristic algorithm. We also evaluate the performance of the proposed algorithm by using the topology of various actual networks.

Nowadays, online learning has become a promising solution to personalize and increase flexibility in the learning-teaching process. However, e-assessment is still questioned in terms of authorship and identity checking. Some virtual learning environments are introducing technological solutions, such as plagiarism detection tools, to increase the security when submitting assessment activities. However, this is a partial solution. When the activities are performed on third-party tools, as it is the case of intelligent tutoring systems, the identity and authorship checking can fail. This paper introduces a modular plagiarism detection tool that combines different input data sources in order to verify the authorship. A case study is presented to show the potential of the tool.

Due to the high number of students enrolled and the relatively small number of available tutors, the assessment of complex assignments is deemed as one of the most critical tasks in Massive Open On-line Courses (MOOCs). Peer assessment is becoming an increasingly popular tool to face this problem and many approaches have been proposed so far to make its outcomes more reliable. A promising approach is FOPA (Fuzzy Ordinal Peer Assessment) that adopts and integrates models coming from Fuzzy Set Theory and Group Decision Making. In this paper we propose a FOPA extension supporting multi-criteria assessment based on rubrics. Students are asked to rank a small number of peer submissions against specified criteria, provided rankings are then transformed in fuzzy preference relations, expanded to obtain missing values and aggregated to establish a global ranking between students’ works with respect to each criterion and globally. The absolute grades of all submissions are then calculated.

Massive Open Online Courses (MOOCs) arose as a way of transcending formal higher education by realizing technology-enhanced formats of learning and instruction and by granting access to a wide audience way beyond students enrolled in any one Higher Education Institution. However, while MOOCs have been reported as an efficient and important educational tool, yet there is a number of issues and problems related to their educational impact. More specifically, there is an important number of drop outs during a course, little participation, and lack of students’ motivation and engagement overall. To overcome these limitations, Conversational pedagogical agents have arisen to guide and support student dialogue using natural language both in individual and collaborative settings. Conversational agents have been produced to meet a wide variety of applications and studies exploring the usage of such agents have led to positive results. Integrating this type of artificial agents into MOOCs is expected to trigger productive peer interaction in discussion groups. In this paper, we present a state-of-the-art study of the use of conversational agents to support collaborative learning in the context of MOOCs. The ultimate goal of this study is to analyze the potential of conversational agents to considerably increase the engagement and the commitment of MOOC students, reducing consequently, the overall MOOCs dropout rate. The research reported in this paper is currently undertaken within the research project colMOOC funded by the European Commission.

The seminars are a learning mechanism that serve to complement regular education with more specialized content. Different learning materials are generated from the seminars activity, such as documentation, presentations, videos, exercises… These materials can also be used for teaching purposes in other contexts. The management and distribution of these materials is normally done by hand or stored in a digital repository. The problem with these tools is that they facilitate storage management but are not focused to support their formative application. This article describes the experience of creating a web tool that allows to take profit of the materials generated in a seminar called Zaragoza Linguistics for later teaching use. The functionality has been implemented using web scraping techniques and the YouTube web services API.

Educational Virtual Worlds (EVWs) provide an interesting and engaging educational platform to promote student learning. Virtual classrooms should be designed to be responsive and adaptive enough to judge the emotional level of the learner and should be capable of providing required support and feed-back to maintain the engagement of the learner to achieve their learning goals. Virtual agents with affective capabilities have potential to improve the effectiveness and efficiency of the e-learning and virtual learning platforms. This research study will explore understanding of learners’ emotional states that evolve during the learning process within a EVW by drawing on theories related to epistemic emotions and develop approaches to recognise and intervene in a timely manner based on the individual’s feelings being experienced at that moment. To derive that emotional state, this paper also presents a multi-data and multimodal approach to judge the underlying emotional state of the learner within EVW and the most appropriate intervention for that learner.

Smart grid (SG) is an efficient electrical grid that provides opportunities to manage the energy load in a reliable and efficient way. Moreover, smart meters (SMs) are introduced, which play a vital role in communication between homes and SG. SMs manage and monitor the energy consumption of homes. SGs and SMs produce a big data, which is very hard to store and process even with cloud computing. For this purpose fog computing concept is introduced, which provides a good environment for computing and storing the data of SGs and SMs before transmitting them to cloud. The concept of fog computing, acts as a bridge between cloud and SGs. In this paper, a cloud-fog based architecture is integrated with SG for efficient energy management of buildings with SMs. To manage the energy requirements of consumers, micro grids (MGs) are available near to the buildings. Fogs are distributed over the world, overhaul the cloud via important features, including low latency and increased security for MGs. To balance the load on cloud and fogs, three load balancing algorithms are used. These algorithms are round robin (RR), throttled and greedy. Closest datacenter policy is used to compare their results. Greedy gives better results than RR and throttled.

Nowadays more and more people or enterprises prefer to outsource their data to the remote cloud servers. However how to efficiently and flexibly share the outsourced data storage secure is a very challenge problem, the cryptographic primitive of functional encryption is such a technique to solve this problem. In this paper, we show several new functional encryption schemes, which are more efficient than the original schemes, by using a novel technique. This technique maybe also has potential applications in other related schemes. Finally we conclude our paper with many interesting open problems.

Due to the limited computational resources of data owners and the developments in cloud computing, more and more data owners choose to store data on the cloud to reduce their own storage burden. When the data owner wants to re-select a new cloud service provider, the data needs to be transmitted from Cloud A to Cloud B. How to ensure that data has been safely transferred to Cloud B and Cloud A honestly implementing data deletion has become a hot topic for many scholars. To solve these problems, Liang et al. proposed a provable data transfer protocol based on provable data possession and deletion for secure cloud storage. However, we find a security flaw in their scheme. In this paper, we first review their scheme and then present our attack in detail. Finally, we propose an improved scheme which can resist the attack well.

Chinese part-of-speech tagging (POS tagging) is a key part of Chinese Natural Language Processing (CNLP) Research. Using POS tagging can better understand semantics and improve the efficiency of Natural Language Processing. This paper proposes a method of POS tagging based on a bidirectional GRU and CRF hybrid model, which can automatically learn features and reduce operational complexity. Under the same conditions, compared with Bi-LSTM+CRF, CNN+LSTM, LSTM+CRF and HMM, the model obtains the best accuracy.

Cloud computing is major component in our daily life; Integration of Cloud with smart grid brings an important role in electricity management. Fog computing concept is also introduced in this paper which helps to minimize the load on cloud. Many techniques are introduced in papers that includes Round Robin (RR), Genetic Algorithm (GA) and Binary Particle Swarm Optimization (BPSO) etc. In this paper authors introduce First Come First Serve (FCFS) load balancing technique with the broker policy of Closest Data Center to allocate resources for Virtual Machines (VM). FCFS algorithm results are compared with existing known algorithms which includes RR and Throttled algorithm. The Response Time (RT) is less in some clusters as compared to RR and Throttled algorithm. The main goal is to optimise the Response Time (RT) on cloud.

This article discusses some existing techniques and methods in data analytics, which aim to identify, extract and assess insights from both unstructured and structured datasets. However, most of them have limitations due to either an over-specialisation or an attempt to provide a general “data panacea”. In this work, we will also suggest a potential approach, which aims to overcome such challenges. In fact, specific enhancements to the existing methods are likely to lead to the creation of a tool that is more suited to rapid decision making in business, through the use of clear visualisations supported by novel machine learning methods.

The paper is focused on the topic of analytics and its connection to strategic information systems. Analytics strategy and IT strategy, however, are not separated areas within an organization – they are mutually connected and moreover, they are connected to the overall business strategy as well as other strategies of functional areas within management. Thus, the paper covers the connection of analytics strategy with IT strategy and the overall business strategies. Based on literature review, there are provided visualizations expressing the placement of analytics strategy to other strategies in management.

This article presents an overview of data inconsistencies and a review of approaches to resolve various levels of data inconsistencies. It provides a discussion of the approaches, which motivates a Bayesian Network approach in inconsistency resolution.

When predicting consumption of Liquefied Petroleum Gas (LPG) it is profitable to know the consumer type. Different consumers, like a single-family house, a bakery or a primary school, require different predictive models to produce qualitative results. Application of meta-learning makes possible an automated approach to LPG consumption prediction through model selection. Additionally, such solution improves the overall prediction quality.

Support vector machine (SVM) is a popular classifier that has been used to solve a broad range of problems. Unfortunately, its applications are limited by computational complexity of training which is $$O(t^3)$$, where t is the number of vectors in the training set. This limitation makes it difficult to find a proper model, especially for non-linear SVMs, where optimization of hyperparameters is needed. Nowadays, when datasets are getting bigger in terms of their size and the number of features, this issue is becoming a relevant limitation. Furthermore, with a growing number of features, there is often a problem that a lot of them may be redundant and noisy which brings down the performance of a classifier. In this paper, we address both of these issues by combining a recursive feature elimination algorithm with our evolutionary method for model and training set selection. With all of these steps, we reduce both the training and classification times of a trained classifier. We also show that the model obtained using this procedure has similar performance to that determined with other algorithms, including grid search. The results are presented over a set of well-known benchmark sets.

Cloud is a pool of virtualized resources. Integrating cloud in a smart grid environment helps to efficiently utilize the energy resources while fulfilling the energy demands of residential users. However, when number of users increase it is difficult to efficiently utilize the cloud resources to handle so many user requests. Fog reduce the latency, processing and response time of user requests. In this paper, cloud-fog based environment for efficient energy management is proposed. The objective of achieving maximum performance is also formulated mathematically in this paper. Simulations in CloudAnalyst are performed to compare and analyze the performance of load balancing algorithms: Round Robin (RR), Throttled, and Weighted Round Robin (WRR) and service broker policies: Service Proximity Policy, Optimize Response Time, Dynamically Reconfigure with Load, and New Dynamic Service Proximity. Simulation results showed that Throttled load balancing algorithm give better response time than RR and WRR.

Breast cancer is the most pervasive form of cancers in women, therefore automated algorithms for cancer detection, and analysis of hematoxylin and eosin stained breast-cancer histopathology images are being actively developed worldwide. In this paper, we propose multi-scale voting classifiers which operate on clinically-relevant features extracted from such images, and apply them to classify real-life breast-cancer data. The extensive experiments (encompassing cross-validation scenarios backed up with statistical tests) showed that our models deliver high-quality classification, can be learned quickly, and offer instant operation.