Computing and Informatics (E-Journal - Institute of Informatics, SAS, Bratislava)
Not a member yet
1506 research outputs found
Sort by
Machine Learning Based Classifier for Service Function Chains
Get PDFUsing service function chains, Internet Service Providers can customize the use of service functions that process the network flows belonging to their customers. Each network flow is injected into a service chain according to the flow features. Since most of the malicious applications try not to get the proper analysis by imitating some valid and famous applications, classification based on simple flow features may waste processing power by using inappropriate service chains for evasive flows. In this paper, we have explored an application-aware classification approach using machine learning methods. Using CatBoost as a machine learning method, a model is created and used for traffic classification. We have provided some statistical reports on how this approach is compared with simple flow feature-based approaches in malicious environments and how feature selection can impact classification correctness. Choosing the most suitable number of features at the right time can beat traditional approaches in classification quality and provide better results in the service function chaining environment
Optimizing Data Placement for Cost Effective and High Available Multi-Cloud Storage
Get PDFWith the advent of big data age, data volume has been changed from trillionbyte to petabyte with incredible speed. Owing to the fact that cloud storage offers the vision of a virtually infinite pool of storage resources, data can be stored and accessed with high scalability and availability. But a single cloud-based data storage has risks like vendor lock-in, privacy leakage, and unavailability. Multi-cloud storage can mitigate these risks with geographically located cloud storage providers. In this storage scheme, one important challenge is how to place a user's data cost-effectively with high availability. In this paper, an architecture for multi-cloud storage is presented. Next, a multi-objective optimization problem is defined to minimize total cost and maximize data availability simultaneously, which can be solved by an approach based on the non-dominated sorting genetic algorithm II (NSGA-II) and obtain a set of non-dominated solutions called the Pareto-optimal set. Then, a method is proposed which is based on the entropy method to determine the most suitable solution for users who cannot choose one from the Pareto-optimal set directly. Finally, the performance of the proposed algorithm is validated by extensive experiments based on real-world multiple cloud storage scenarios
Overlapping Community Detection Extended from Disjoint Community Structure
Get PDFCommunity detection is a hot issue in the study of complex networks. Many community detection algorithms have been put forward in different fields. But most of the existing community detection algorithms are used to find disjoint community structure. In order to make full use of the disjoint community detection algorithms to adapt to the new demand of overlapping community detection, this paper proposes an overlapping community detection algorithm extended from disjoint community structure by selecting overlapping nodes (ONS-OCD). In the algorithm, disjoint community structure with high qualities is firstly taken as input, then, potential members of each community are identified. Overlapping nodes are determined according to the node contribution to the community. Finally, adding overlapping nodes to all communities they belong to and get the final overlapping community structure. ONS-OCD algorithm reduces the computation of judging overlapping nodes by narrowing the scope of the potential member nodes of each community. Experimental results both on synthetic and real networks show that the community detection quality of ONS-OCD algorithm is better than several other representative overlapping community detection algorithms
Information Technology of Generalized Model Creation of Complex Technical Objects
Get PDFThe paper introduces a knowledge representation framework for design and geometrical modelling of complex technical objects such as ships, aircrafts, cars, etc. The design process cannot be fully automated yet because of a lot of technical and economical factors that influence the decisions during that process. In order to make the process more efficient, a knowledge modelling framework is suggested. The basic principles of conceptual knowledge modelling and data exchange framework are presented. A practical use case of aircraft ramp modelling is provided
Measuring Sentences Similarity Based on Discourse Representation Structure
Get PDFThe problem of measuring similarity between sentences is crucial for many applications in Natural Language Processing (NLP). Most of the proposed approaches depend on similarity of words in sentences. This research considers semantic relations between words in calculating sentence similarity. This paper uses Discourse Representation Structure (DRS) of natural language sentences to measure similarity. DRS captures the structure and semantic information of sentences. Moreover, the estimation of similarity between sentences depends on semantic coverage of relations of the first sentence in the other sentence. Experiments show that exploiting structural information achieves better results than traditional word-to-word approaches. Moreover, the proposed method outperforms similar approaches on a standard benchmark dataset
Reverse Intervention for Dealing with Malicious Information in Online Social Networks
Get PDFMalicious information is often hidden in the massive data flow of online social networks. In “We Media'' era, if the system is closed without intervention, malicious information may spread to the entire network quickly, which would cause severe economic and political losses. This paper adopts a reverse intervention strategy from the perspective of topology control, so that the spread of malicious information could be suppressed at a minimum cost. Noting that as the information spreads, social networks often present a community structure and multiple malicious information promoters may appear. Therefore, this paper adopts a divide and conquer strategy and proposes an intervention algorithm based on subgraph partitioning, in which we search for some influential nodes to block or release clarification. The main algorithm consists of two main phases. Firstly, a subgraph partitioning method based on community structure is given to quickly extract the community structure of the information dissemination network. Secondly, a node blocking and clarification publishing algorithm based on the Jordan Center is proposed in the obtained subgraphs. Experiments show that the proposed algorithm can effectively suppress the spread of malicious information with a low time complexity compared with the benchmark algorithms
Case Study on Human-Robot Interaction of the Remote-Controlled Service Robot for Elderly and Disabled Care
Get PDFThe tendency of continuous aging of the population and the increasing number of people with mobility difficulties leads to increased research in the field of Assistive Service Robotics. These robots can help with daily life tasks such as reminding to take medications, serving food and drinks, controlling home appliances and even monitoring health status. When talking about assisting people in their homes, it should be noted that they will, most of the time, have to communicate with the robot themselves and be able to manage it so that they can get the most out of the robot's services. This research is focused on different methods of remote control of a mobile robot equipped with robotic manipulator. The research investigates in detail methods based on control via gestures, voice commands, and web-based graphical user interface. The capabilities of these methods for Human-Robot Interaction (HRI) have been explored in terms of usability. In this paper, we introduce a new version of the robot Robco 19, new leap motion sensor control of the robot and a new multi-channel control system. The paper presents methodology for performing the HRI experiments from human perception and summarizes the results in applications of the investigated remote control methods in real life scenarios
Automatic Query Refining Based on Eye-Tracking Feedback
Get PDFThis paper presents a new method named AQueReBET, which automatically refines a query set by an information seeker searching on the web. A revelation of the intention of an information seeker who is running a search can bring a significant improvement to the search process, and to browsing as well. It is practically impossible to acquire such intention by the explicit indication (feedback) due to the fact that web browsing takes place in real time. Therefore the intention must be determined in some other way. We hypothesize that it can be approximated by means of the implicit feedback preferably in the form of data from an eye tracker and mouse. We propose a method which automatically refines a seeker’s search query and thus we can offer documents with higher relevance, decrease the number of query reformulations and increase the seeker’s satisfaction. The query refinement is based on an analysis of gaze data from an eye tracker and also on groupization. In the proposed method, we calculate word-level importance based on term frequency, term uniqueness (tf-idf) and total fixation duration within the subdocument (word's snippet in search results)
Generalized Selection Method
Get PDFIn this paper we introduce new selection method, 3-selection method. This method tries to generalize the most used selection methods in Genetic Algorithms (GA). Our new method involves both proportional and rank-based methods (order-based) and, moreover, it allows scaling of selection pressure with higher precision. This method is based on defining the shape of probability density distribution which is adjustable by parameters of our method. In addition, our method has one more attribute which adds randomness of selection
Improved Multi-Population Differential Evolution for Large-Scale Global Optimization
Get PDFDifferential evolution (DE) is an efficient population-based search algorithm with good robustness, however, it is challenged to deal with high-dimensional problems. In this paper, we propose an improved multi-population differential evolution with best-and-current mutation strategy (mDE-bcM). The population is divided into three subpopulations based on the fitness values, each of subpopulations uses different mutation strategy. After crossover, mutation and selection, all subpopulations are updated based on the new fitness values of their individuals. An improved mutation strategy is proposed, which uses a new approach to generate base vector that is composed of the best individual and current individual. The performance of mDE-bcM is evaluated on a set of 19 large-scale continuous optimization problems, a comparative study is carried out with other state-of-the-art optimization techniques. The results show that mDE-bcM has a competitive performance compared to the contestant algorithms and better efficiency for large-scale optimization problems