Кафедра авіаційних комп'ютерно-інтегрованих комплексів (НОВА)

Permanent URI for this communityhttp://er.nau.edu.ua/handle/NAU/58724

News

Відповідальний за розділ: Провідний фахівець кафедри авіаційних комп'ютерно-інтегрованих комплексів інституту інформаційно-діагностичних систем Шугалєй Людмила Петрівна. E-mail: shugaley2005@ukr.net

Browse

Browsing Кафедра авіаційних комп'ютерно-інтегрованих комплексів (НОВА) by Subject "004.85(045)"

Filter results by typing the first few letters
Now showing 1 - 2 of 2
  • No Thumbnail Available
    Item
    Automated Adjustment System of Restricted Boltzmann Machine
    (Київ «Освіта України», 2019-06) Sineglazov, V. M.; Tofaniuk, O. R.
    In this paper the problem of learning the deep believe neural network with help of a restricted Boltzmann machine and the choose of an optimal algorithm for its training is considered. Different algorithms of restricted Boltzmann machine training, which are used for the pre-training of deep believe neural network, are considered, in order to increase the efficiency of this network and further solve the problem of structural-parametric synthesis of deep believe neural network. This task represents the task of justifying the necessity of optimal choice of the restricted Boltzmann machine adjustment algorithm for improving the quality of training of the neural network. To solve this problem, it is suggested to create an automated adjustment system of restricted Boltzmann machine, which choose the optimal training algorithm for this neural network
  • No Thumbnail Available
    Item
    Deep Learning Fuzzy Classifier
    (Київ «Освіта України», 2019-06) Sineglazov, V. M.; Koniushenko, R. S.
    It is considered a classification problem solution based on analysys of represented review. It’s shown that the neural networks has important advantages beside other methods, such as: classification using the nearest neighbor method, support vector classification, classification using decision trees, etc. Amoun of artifisial neural networks exists futher networks have the simplest structure, but the precission of the solution can be increased with help of deep learning approache, which is supposed the use of additional neural network for the solution of pretraining tasks(deep believe networks). It’s proposed new tophology wich consist of: Takagi-Sugeno-Kang fuzzy classifier and Limited Boltzmann Machine neural network. Despite on this thopology was proposed early in this article it’s carried out enough researches that permited to specify the learning algorithm. An example of proposed algorithm implantation is represented