Publikacje
Filtry
wszystkich: 844
Katalog Publikacji
Rok 2017
-
Minimizing Distribution and Data Loading Overheads in Parallel Training of DNN Acoustic Models with Frequent Parameter Averaging
PublikacjaIn the paper we investigate the performance of parallel deep neural network training with parameter averaging for acoustic modeling in Kaldi, a popular automatic speech recognition toolkit. We describe experiments based on training a recurrent neural network with 4 layers of 800 LSTM hidden states on a 100-hour corpora of annotated Polish speech data. We propose a MPI-based modification of the training program which minimizes the...
-
Mobile devices and computing cloud resources allocation for interactive applications
PublikacjaUsing mobile devices such as smartphones or iPads for various interactive applications is currently very common. In the case of complex applications, e.g. chess games, the capabilities of these devices are insufficient to run the application in real time. One of the solutions is to use cloud computing. However, there is an optimization problem of mobile device and cloud resources allocation. An iterative heuristic algorithm for...
-
Modeling and Simulation for Exploring Power/Time Trade-off of Parallel Deep Neural Network Training
PublikacjaIn the paper we tackle bi-objective execution time and power consumption optimization problem concerning execution of parallel applications. We propose using a discrete-event simulation environment for exploring this power/time trade-off in the form of a Pareto front. The solution is verified by a case study based on a real deep neural network training application for automatic speech recognition. A simulation lasting over 2 hours...
-
Multi-agent large-scale parallel crowd simulation
PublikacjaThis paper presents design, implementation and performance results of a new modular, parallel, agent-based and large scale crowd simulation environment. A parallel application, implemented with C and MPI, was implemented and run in this parallel environment for simulation and visualization of an evacuation scenario at Gdansk University of Technology, Poland and further in the area of districts of Gdansk. The application uses a...
-
Multidisciplinary Open System Transferring Knowledge for R2B Development
PublikacjaDespite many efforts, there is still a serious problem in transferring knowledge from research to business. The problem is especially visible in Poland – the cooperation of R2B is ineffective. We are trying to solve this problem using some IT support. The manuscript presents some solutions developed at the Gdańsk University of Technology. In particular, the platform called “MOST Knowledge” is deeply described. Its layer architecture,...
-
Optymalizacja strategii sieci inteligentnych agentów za pomocą programowania genetycznego w systemie rozproszonym realizującym paradygmat volunteer computing
PublikacjaDynamicznie rosnąca złożoność i wymagania w odniesieniu do rozproszonych systemów informatycznych utrudnia zarządzanie dostępnymi zasobami sprzętowymi i programistycznymi. Z tego powodu celem rozprawy jest opracowanie wielokryterialnej metody programowania genetycznego, która pozwala na optymalizację strategii zespołu inteligentnych agentów programistycznych w zakresie zarządzania systemem realizującym paradygmat volunteer computing....
-
Parallelization of Selected Algorithms on Multi-core CPUs, a Cluster and in a Hybrid CPU+Xeon Phi Environment
PublikacjaIn the paper we present parallel implementations as well as execution times and speed-ups of three different algorithms run in various environments such as on a workstation with multi-core CPUs and a cluster. The parallel codes, implementing the master-slave model in C+MPI, differ in computation to communication ratios. The considered problems include: a genetic algorithm with various ratios of master processing time to communication...
-
Path-based methods on categorical structures for conceptual representation of wikipedia articles
PublikacjaMachine learning algorithms applied to text categorization mostly employ the Bag of Words (BoW) representation to describe the content of the documents. This method has been successfully used in many applications, but it is known to have several limitations. One way of improving text representation is usage of Wikipedia as the lexical knowledge base – an approach that has already shown promising results in many research studies....
-
Performance evaluation of unified memory and dynamic parallelism for selected parallel CUDA applications
PublikacjaThe aim of this paper is to evaluate performance of new CUDA mechanisms—unified memory and dynamic parallelism for real parallel applications compared to standard CUDA API versions. In order to gain insight into performance of these mechanisms, we decided to implement three applications with control and data flow typical of SPMD, geometric SPMD and divide-and-conquer schemes, which were then used for tests and experiments. Specifically,...
-
Platforma IPMed jako elastyczne źródło danych dla medycznych zespołów naukowych
PublikacjaPrzedstawiono architekturę rozproszonej platformy IPMed, która umożliwia akwizycję i przechowywanie zanonimizowanych danych medycznych. Pokazano wyniki zastosowania platformy w obszarze badań z zakresu hemodynamiki układu krążenia. Uzyskane dane pozwoliły na weryfikację hipotezy zespołu badawczego i określenie reguł rekomendacji w z leczeniu udarów.
-
Practical Evaluation of Internet Systems' Security Mechanisms
PublikacjaA proposed Internet systems security layer with context-oriented security mechanisms reduces the risk associated with possible vulnerabilities. A metric of the system trust level is proposed, and then evaluated according to a university Internet system.
-
Qualitative evaluation of distributed clinical systems supporting research teams working on large-scale data
PublikacjaInthispaper,fivecontemporaryscalablesystemstosupportmedicalresearchteams are presented. Their functionalities extend from heterogeneous unstructured data acquisition through large-scale data storing, to on-the-fly analyzing by using robust methods. Such kinds of systems can be useful in the development of new medical procedures and recommendation rules for decision support systems. A short description of each of them is provided....
-
SERVICE ORIENTED SCENARIOS FOR SUPPORT ENTERPRISE PROJECT MANAGEMENT
PublikacjaIn the paper, a service oriented architecture is used to describe various scenarios of human-oriented enterprises. They can be modeled by BPMN to show basic information flow. We consider business scenarios to present another modeling solution called service orchestration or service mashup. That approach gives an opportunity for the right management through taking the best possible choice of services, or composition of services....
-
Smart Services supporting Drivers in Effective Cars Parking
PublikacjaThe paper shows how to design and implement intelligent, service oriented systems for management of free places in distributed car parks. The idea of SOA technology is widely used. It is assumed that a state of every car park is automatically determined by a special electronic infrastructure including IoT solutions. The information received from such an infrastructure is sent to the main management system, which processes it and,...
-
Spectral Clustering Wikipedia Keyword-Based search Results
PublikacjaThe paper summarizes our research in the area of unsupervised categorization of Wikipedia articles. As a practical result of our research, we present an application of spectral clustering algorithm used for grouping Wikipedia search results. The main contribution of the paper is a representation method for Wikipedia articles that has been based on combination of words and links and used for categoriation of search results in this...
-
Towards facts extraction from text in Polish language
PublikacjaNatural Language Processing (NLP) finds many usages in different fields of endeavor. Many tools exists allowing analysis of English language. For Polish language the situation is different as the language itself is more complicated. In this paper we show differences between NLP of Polish and English language. Existing solutions are presented and TEAMS software for facts extraction is described. The paper shows also evaluation of...
-
Uczenie maszynowe do samoorganizacji systemów rozproszonych w zastosowaniach gospodarczych
PublikacjaW pracy omówiono uczenie maszynowe do samoorganizacji systemów rozproszonych w zastosowaniach gospodarczych ze szczególnym uwzględnieniem sieci neuronowych do predykcji finansowych oraz szacowania ratingu przedsiębiorstw. Oprócz sieci neuronowych, istotną rolę w przygotowaniu i testowaniu informatycznych systemów finansowych może pełnić programowanie genetyczne. Z tego powodu omówiono uczenie maszynowe w aplikacjach konstruowanych...
-
Using Redis supported by NVRAM in HPC applications
PublikacjaNowadays, the efficiency of storage systems is a bottleneck in many modern HPC clusters. High performance in the traditional approach – processing using files – is often difficult to obtain because of a model’s complexity and its read/write patterns. An alternative approach is to apply a key-value database, which usually has low latency and scales well. On the other hand, many key-value stores suffer from a limitation of memory...
-
WordVenture - COOPERATIVE WordNet EDITOR Architecture for Lexical Semantic Acquisition
PublikacjaThis article presents architecture for acquiring lexical semantics in a collaborative approach paradigm. The system enables functionality for editing semantic networks in a wikipedia-like style. The core of the system is a user-friendly interface based on interactive graph navigation. It has been used for semantic network presentation, and brings simultaneously modification functionality.
Rok 2016
-
A Fail-Safe NVRAM Based Mechanism for Efficient Creation and Recovery of Data Copies in Parallel MPI Applications
PublikacjaThe paper presents a fail-safe NVRAM based mechanism for creation and recovery of data copies during parallel MPI application runtime. Specifically, we target a cluster environment in which each node has an NVRAM installed in it. Our previously developed extension to the MPI I/O API can take advantage of NVRAM regions in order to provide an NVRAM based cache like mechanism to significantly speed up I/O operations and allow to preload...
-
A Parallel MPI I/O Solution Supported by Byte-addressable Non-volatile RAM Distributed Cache
PublikacjaWhile many scientific, large-scale applications are data-intensive, fast and efficient I/O operations have become of key importance for HPC environments. We propose an MPI I/O extension based on in-system distributed cache with data located in Non-volatile Random Access Memory (NVRAM) available in each cluster node. The presented architecture makes effective use of NVRAM properties such as persistence and byte-level access behind...
-
Adaption of Open Science Paradigm at Gdańsk University of Technology
PublikacjaThe paper describes the basic principles standing behind Open Access and will try to dispel the doubts behind them. Further the solution adapted by the Gdańsk University of Technology is presented. Later on the systems that the University uses and means of their integration into a complex Open Science platform are described. Finally the procedures needed to publish a document in the repository alongside the University's policy...
-
Alternative Modeling of Volunteer Based Systems
PublikacjaThe canonical approach to model a flow of a voluntary based computing, as presented in Chapter 7 may tax the capacity of the simulator, especially when a more complex model of Voluntary Computing, proposed in [1], is used which would imply that, for typical heterogeneous groups of volunteers, several thousands of threads would have to be run inside the simulator, to mimic closely the behaviour of every volunteer. This chapter presents...
-
APIS – Agent Platform for Integration of Services
PublikacjaThe paper presents an approach to creating the platform for development and evaluation of complex tasks execution algorithms. Proposed solution is based on an agent paradigm where independent peers can cooperate and negotiate in order to execute specified tasks which are defined only by the final output description. Tasks are realized by the means of services exposed by different peers. In case when there is no single service fulfilling...
-
Application of Web-GIS and Cloud Computing to Automatic Satellite Image Correction
PublikacjaRadiometric calibration of satellite imagery requires coupling of atmospheric and topographic parameters, which constitutes serious computational problems in particular in complex geographical terrain. Successful application of topographic normalization algorithms for calibration purposes requires integration of several types of high-resolution geographic datasets and their processing in a common context. This paper presents the...
-
Automatic Detection of Nerves in Confocal Corneal Images with Orientation-Based Edge Merging
PublikacjaThe paper presents an algorithm for improving results of automatic nerve detections in confocal microscopy images of human corneal. The method is designed as a postprocessing step of regular detection. After the nerves are initially detected, the algorithms attempts to improve the results by filling unde-sired gaps between single nerves detections in order to correctly mark the entire nerve instead of only parts of it. This approach...
-
Automatic Discovery of IaaS Cloud Workload Types
PublikacjaThe paper presents an approach to automatic discovery of workloads types. We perform functional characteristics of the workloads executed in our cloud environment, that have been used to create model of the computations. To categorize the resources utilization we used K-means algorithm, that allow us automatically select six types of computations. We perform analysis of the discovered types against to typical computational benchmarks,...
-
Benchmarking Performance of a Hybrid Intel Xeon/Xeon Phi System for Parallel Computation of Similarity Measures Between Large Vectors
PublikacjaThe paper deals with parallelization of computing similarity measures between large vectors. Such computations are important components within many applications and consequently are of high importance. Rather than focusing on optimization of the algorithm itself, assuming specific measures, the paper assumes a general scheme for finding similarity measures for all pairs of vectors and investigates optimizations for scalability...
-
Category-Based Workload Modeling for Hardware Load Prediction in Heterogeneous IaaS Cloud
PublikacjaThe paper presents a method of hardware load prediction using workload models based on application categories and high-level characteristics. Application of the method to the problem of optimization of virtual machine scheduling in a heterogeneous Infrastructure as a Service (IaaS) computing cloud is described.
-
Checkpointing of Parallel MPI Applications using MPI One-sided API with Support for Byte-addressable Non-volatile RAM
PublikacjaThe increasing size of computational clusters results in an increasing probability of failures, which in turn requires application checkpointing in order to survive those failures. Traditional checkpointing requires data to be copied from application memory into persistent storage medium, which increases application execution time as it is usually done in a separate step. In this paper we propose to use emerging byte-addressable...
-
Clustering Context Items into User Trust Levels
PublikacjaAn innovative trust-based security model for Internet systems is proposed. The TCoRBAC model operates on user profiles built on the history of user with system interaction in conjunction with multi-dimensional context information. There is proposed a method of transforming the high number of possible context value variants into several user trust levels. The transformation implements Hierarchical Agglomerative Clustering strategy....
-
Completeness and Consistency of the System Requirement Specification
PublikacjaAlthough the System Requirement Specification, as a first formal and detailed document, is the base for the software project in classic software methodologies, there is a noticeable problem of assuring the completeness of this document. The lack of its completeness causes uncertainty of the project foundations. This was one of motivations for agile methodologies – if the SRS cannot be easily validated, if it can change in late project...
-
Considerations of Computational Efficiency in Volunteer and Cluster Computing
PublikacjaIn the paper we focus on analysis of performance and power consumption statistics for two modern environments used for computing – volunteer and cluster based systems. The former integrate computational power donated by volunteers from their own locations, often towards social oriented or targeted initiatives, be it of medical, mathematical or space nature. The latter is meant for high performance computing and is typically installed...
-
Depth Images Filtering In Distributed Streaming
PublikacjaIn this paper, we propose a distributed system for point cloud processing and transferring them via computer network regarding to effectiveness-related requirements. We discuss the comparison of point cloud filters focusing on their usage for streaming optimization. For the filtering step of the stream pipeline processing we evaluate four filters: Voxel Grid, Radial Outliner Remover, Statistical Outlier Removal and Pass Through....
-
DEPTH IMAGES FILTERING IN DISTRIBUTED STREAMING
PublikacjaIn this paper we discuss the comparison of point cloud filters focusing on their applicability for streaming optimization. For the filtering stage within a stream pipeline processing we evaluate three filters: Voxel Grid, Pass Through and Statistical Outlier Removal. For the filters we perform series of the tests aiming at evaluation of changes of point cloud size and transmitting frequency (various fps ratio). We propose a distributed...
-
Energy Consumption Modeling in SPMD and DAC Applications
PublikacjaIn this chapter, we show a study of energy consumption during execution of SPMD and DAC application – the same applications which time of execution we modeled in the previous two chapters. We measured an average power usage at a single node of the GALERA+ cluster during application execution and then we modeled the total energy consumption by the application. Next we simulated the applications using MERPSYS and we compared the...
-
Executing Multiple Simulations in the MERPSYS Environment
PublikacjaThe chapter investigates the steps necessary to perform a simulation instance in the MERPSYS environment and discusses potential limitations in case when vast numbers of simulations are required. An extended architecture is proposed which includes a JMS-based simulation queue and multiple distributed simulators, overcoming the potential bottlenecks. The chapter introduces also methods for preparing suites of multiple simulations...
-
Game with a Purpose for Mappings Verification
PublikacjaMappings verification is a laborious task. The paper presents a Game with a Purpose based system for verification of automatically generated mappings. General description of idea standing behind the games with the purpose is given. Description of TGame system, a 2D platform mobile game with verification process included in the gameplay, is provided. Additional mechanisms for anti-cheating, increasing player’s motivation and gathering...
-
Harmony Search for Data Mining with Big Data
PublikacjaIn this paper, some harmony search algorithms have been proposed for data mining with big data. Three areas of big data processing have been studied to apply new metaheuristics. The first problem is related to MapReduce architecture that can be supported by a team of harmony search agents in grid infrastructure. The second dilemma involves development of harmony search in preprocessing of data series before data mining. Moreover,...
-
Harmony Search for Self-configuration of Fault–Tolerant and Intelligent Grids
PublikacjaIn this paper, harmony search algorithms have been proposed to self-configuration of fault-tolerant grids for big data processing. Self-configuration of computer grids lies in the fact that new computer nodes are automatically configured by software agents and then integrated into the grid. A base node works due to several configuration parameters that define some aspects of data communications and energy power consumption. We...
-
Harmony Search to Self-Configuration of Fault-Tolerant Grids for Big Data
PublikacjaIn this paper, harmony search algorithms have been proposed to self-configuration of fault-tolerant grids for big data processing. Some tasks related to big data processing have been considered. Moreover, two criteria have been applied to evaluate quality of grids. The first criterion is a probability that all tasks meet their deadlines and the second one is grid reliability. Furthermore, some intelligent agents based on harmony...
-
Identification of category associations using a multilabel classifier
PublikacjaDescription of the data using categories allows one to describe it on a higher abstraction level. In this way, we can operate on aggregated groups of the information, allowing one to see relationships that do not appear explicit when we analyze the individual objects separately. In this paper we present automatic identification of the associations between categories used for organization of the textual data. As experimental data...
-
Ischemic stroke in general practice - risk factors, prognosis - prospective single center study
PublikacjaBackground. Ischemic stroke is one of the leading causes of death and disability in Poland, in the European Union and in the United States. Awareness related to the risk factors, particularly the modifiable ones, is of crucial importance in general practice and may improve early and long-term prognoses. Objectives. The aim of this study was to assess the risk factors and half-year survival rate in patients with ischemic stroke...
-
KernelHive: a new workflow-based framework for multilevel high performance computing using clusters and workstations with CPUs and GPUs
PublikacjaThe paper presents a new open-source framework called KernelHive for multilevel parallelization of computations among various clusters, cluster nodes, and finally, among both CPUs and GPUs for a particular application. An application is modeled as an acyclic directed graph with a possibility to run nodes in parallel and automatic expansion of nodes (called node unrolling) depending on the number of computation units available....
-
Metody neuronowe do prognozowania finansowego
PublikacjaSztuczne sieci neuronowe mogą być stosowane do prognozowania kursów akcji na giełdzie, oceny wiarygodności kredytobiorców czy prognozowania kryzysów bankowych. W referacie omówiono zasady współpracy sieci neuronowych z algorytmami ewolucyjnymi oraz metodą wektorów wspierających. Ponadto, odniesiono się do pozostałych metod sztucznej inteligencji, które stosowane są w finansach.
-
Modeling DAC Application Execution Time
PublikacjaAn application written in the Divide And Conquer paradigm is more difficult to model than SPMD application because of complex algorithm, causing use of many coefficients in a computational complexity function. Processors are divided into various layers, each layer contains different number of processors. Data packets processed in different layers and transferred between layers have different length. Moreover first layer processors use...
-
Modeling energy consumption of parallel applications
PublikacjaThe paper presents modeling and simulation of energy consumption of two types of parallel applications: geometric Single Program Multiple Data (SPMD) and divide-and-conquer (DAC). Simulation is performed in a new MERPSYS environment. Model of an application uses the Java language with extension representing message exchange between processes working in parallel. Simulation is performed by running threads representing distinct process...
-
Modeling of Performance, Reliability and Energy Efficiency in Large-Scale Computational Environment
PublikacjaLarge scale of complexity of distributed computational systems imposes special challanges for prediction of quality in such systems.Existing quality models for lower-scale systems include functionality,performance,reliability,flexibility and usability.Among these attributes,performance and reliability have a particular significance to the large-scale systems computing quality modeling due to their strong dependence on the system...
-
Modeling Parallel Applications in the MERPSYS Environment
PublikacjaThe chapter presents how to model parallel computational applications for which simulation of execution in a large-scale parallel or distributed environment is performed within the MERPSYS environment. Specifically, it is shown what approaches can be adopted to model key paradigms often used for parallel applications: master-slave, geometric parallelism (single program multiple data), pipelined and divide-and-conquer applications....
-
Modeling SPMD Application Execution Time
PublikacjaParallel applications in a Single Process Multiple Data paradigm assume splitting huge amounts of data to multiple processors working in parallel at small data packets. As the individual data packets are not independent, the processors must interact with each other to exchange results of the calculations with their adjacent partners and take these results into account in their own computations. An example of SPMD is geometric parallelism...