Computer Science Journal (AGH University of Science and Technology, Krakow)
Not a member yet
476 research outputs found
Sort by
Performance Evaluation Tools for Message-Passing Parallel Programs
No full textThis article presents issues related to the analysis of the quality of parallel programs based on the paradigm of sending messages (including PVM [1], MPI [2]) and the construction of tools enabling this analysis. Information about existing tools is dispersed and usually unilaterally presented by the authors, hence the purpose of this article was to collect this information in one place and to carry out a systematic assessment of tools: their advantages and limitations. This should help you choose the right tool. The summary, based on the comparative analysis presented in the paper, proposes further actions aimed at improving the functionality of the tool
Intelligent System in some Decision Classes’ Environment
No full textAt the end of our century, we can observe revolution in the field of information processing. We are currently witnessing an increase in the amount of data that human beings use in their daily lives, as well as an increase in the possibilities of data processing. We can easily say that the above phenomena are not only dependent on each other, but also interact with each other. In a situation of rapid growth in both the amount of information and computing power, the question may be asked how to improve our perception of reality by having such resources at our disposal. Until now, in the process of learning about the world around him, he used methods whose bases were developed at the end of the last century, thanks to which he had numerical procedures that could be used, among others, in such areas as speech recognition, image recognition to control systems in certain narrowly specialized classes of issues. Notwithstanding this, artificial intelligence research was carried out in many scientific laboratories. Thanks to the results of this work, it was possible to create various types of expert and predictive systems that were based on symbolic calculus, or used numerical techniques. Restrictions on this type of system may include: the use of established numerical procedures, small opportunities to implement the learning process, or rather self-learning, as well as the lack of a universal system architecture to solve problems in various fields. In this way we come to the issues that will be the subject of this article
HiPHET: A Hybrid Approach to Translate Code Mixed Language (Hinglish) to Pure Languages (Hindi and English)
Get PDFBilingual code mixed (hybrid) languages has become very popular in India as a result of the spread of Western technology in the form of the television, the Internet and social media. Due to this increase in usage of code-mixed languages in day-to-day communication, the need for maintaining the integrity of Indian languages has arisen. As a result of this need the tool named Hinglish to Pure Hindi and English Translator was developed. The tool translated in three ways, namely, Hinglish to Pure Hindi and Pure English, Pure Hindi to Pure English and vice versa. The tool has achieved accuracy of 91% in giving Hindi sentences as output and of 84% in giving English sentences as output, where the input sentences were in Hinglish. The tool has also been compared with another similar tool in the paper
Forecasting currency exchange rate time series with fireworks-algorithm-based higher order neural network with special attention to training data enrichment
Get PDFExchange rates are highly fluctuating by nature, thus difficult to forecast. Artificial neural networks (ANN) have proved to be better than statistical methods. Inadequate training data may lead the model to reach suboptimal solution resulting, poor accuracy as ANN-based forecasts are data driven. To enhance forecasting accuracy, we suggests a method of enriching training dataset through exploring and incorporating of virtual data points (VDPs) by an evolutionary method called as fireworks algorithm trained functional link artificial neural network (FWA-FLN). The model maintains the correlation between the current and past data, especially at the oscillation point on the time series. The exploring of a VDP and forecast of the succeeding term go consecutively by the FWA-FLN. Real exchange rate time series are used to train and validate the proposed model. The efficiency of the proposed technique is related to other models trained similarly and produces far better prediction accuracy
A hybrid CNN-LiGRU acoustic modeling using raw waveform sincnet for Hindi ASR
No full textDeep Neural Network (DNN) is currently playing the most vital role in Automatic Speech Recognition (ASR). Convolution Neural Network (CNN) and Recurrent Neural Network (RNN) are the advanced versions of DNN. CNN and RNN are right to deal with spatial and temporal properties of the speech signal, respectively, and both properties have a higher impact on accuracy. In today’s scenario, many acoustic modeling techniques often switches due to the battle of CNNs and RNNs. In the last few years, CNN, with raw speech signal, shows their superiority over precomputed acoustic features. Recently, a novel first convolution layer named SincNet was proposed to produce the interpretable filters with better accuracy. In this work, we proposed a hybrid SincNet-CNN-RNN architecture with low computation cost and high accuracy. Different configurations of the hybrid model were extensively examined to achieve this goal. All experiments were performed on the Hindi speech dataset
FORWARD AND BACKWARD STATIC ANALYSIS FOR CRITICAL NUMERICAL ACCURACY IN FLOATING POINT PROGRAMS
Get PDFIn this article, we introduce a new static analysis for numerical accuracy. Weaddress the problem of determining the minimal accuracy on the inputs and on the intermediary results of a program containing foating-point computations in order to ensure a desired accuracy on the outputs. The main approach is to combine a forward and a backward static analysis, done by abstract interpretation. The backward analysis computes the minimal accuracy needed for the inputs and intermediary results of the program in order to ensure a desired accuracy on the results, specied by the user. In practice, the information collected by our analysis may help to optimize the formats used to represent the values stored in the variables of the program or to select the appropriate sensors. To illustrate our analysis, we have shown a prototype example with experimental results
Exploring convolutional auto-encoders for representation learning on networks
Get PDFA multitude of important real-world or synthetic systems possess network structure. Extending learning techniques such as neural networks to process such non-euclidean data is therefore an important direction for machine learning research. However, till very recently this domain has received comparatively low levels of attention. There is no straight forward application of machine learning to network data as machine learning tools are designed for data, simple euclidean data or grids. To address this challenge the technical focus of this dissertation is on use of graph neural networks for Network Representation Learning (NRL) i.e. learning vector representations of nodes in networks. Learning vector embeddings of graph-structured data is similar to embedding complex data into low-dimensional geometries. After the embedding process is completed, drawbacks associated with graph structured data are overcome. The current inquiry proposes two deep learning auto-encoder based approaches for generating node embeddings. The drawbacks in existing auto-encoder approaches such as shallow architectures and excessive parameters are tackled in the proposed architectures using fully convolutional layers. Extensive experiments are performed on publicly available benchmark network data-sets to highlight the validity of this approach
Study of OpenCL processing models for FPGA device
Get PDFIn our study, we present the results of the implementation of SHA-512 algorithm in FPGA. The distinguished element of our work is that we conducted the work using OpenCL for FPGA which is a relatively new development method for reconfigurable logic. We examine the loop unrolling; as the OpenCL performance optimisation method, and compare the efficiency of the different kernel implementation types: NDRange, Single-Work Item, and SIMD kernels. In conclusions, we compare metrics of the created FPGA accelerator to the corresponding GPGPU solutions. Also, our paper is accompanied by the source code repository to allow the reader to follow and extend our survey
Evolutionary Multi-Agent System with Crowding Factor and Mass Center mechanisms for Multiobjective Optimisation
Get PDFThis work presents some additional mechanisms for Evolutionary Multi-Agent Systems for Multiobjective Optimisation trying to solve problems with population stagnation and loss of diversity. Those mechanisms reward solutions located in a less crowded neighborhood and on edges of the frontier. Both techniques have been described and also some preliminary results have been shown
The Multi-Constrained Multicast Routing Improved by Hybrid Bacteria Foraging-Particle Swarm Optimization
No full textTo solve multicast routing under multiple constraints, it is required to generate a multicast tree that ranges from a source to the destinations with minimum cost subject to several constraints. In this paper, PSO has been embedded with BFO to improve the convergence speed and avoid premature convergence that will be used for solving QoS multicast routing problem. The algorithm proposed here generates a set of delay compelled links to every destination present in the multicast group. Then the Bacteria Foraging Algorithm (BFA) selects the paths to all the destinations sensibly from the set of least delay paths to construct a multicast tree. The robustness of the algorithm being proposed had been established through the simulation. The efficiency and effectiveness of the algorithm being proposed was validated through the comparison study with other existing meta-heuristic algorithms. It shows that our proposed algorithm IBF-PSO outperforms its competitive algorithms