Informacja

Drogi użytkowniku, aplikacja do prawidłowego działania wymaga obsługi JavaScript. Proszę włącz obsługę JavaScript w Twojej przeglądarce.

English (EN)·Polski (PL)·Yкраїнський (UK)
Przejdź do strony domowej biblioteki Centralna Biblioteka Wojskowa Link otwiera się w nowym oknie Logo biblioteki Prolib Integro - strona głównaCentralna Biblioteka Wojskowa

Wyszukujesz frazę "Neural networks" wg kryterium: Temat

  Lista filtrów

Dostawca treści

Wyrównaj
Wyświetlanie 1-15 z 37
Tytuł:
From Linear Classifier to Convolutional Neural Network for Hand Pose Recognition
Autorzy:
Rościszewski, P.
Powiązania:
https://bibliotekanauki.pl/articles/305776.pdf
Data publikacji:
2017
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Tematy:
machine learning
artificial neural networks
computer vision
Opis:
Recently gathered image datasets and new capabilities of high performance computing systems allowed developing new artificial neural network models and training algorithms. Using the new machine learning models, computer vision tasks can be accomplished based on the raw values of image pixels, instead of specific features. The principle of operation of deep artificial neural networks is more and more resembling of what we believe to be happening in the human visual cortex. In this paper we build up an understanding of convolutional neural networks through investigating supervised machine learning methods suchas K-Nearest Neighbors, linear classifiers and fully connected neural networks. We provide examples and accuracy results based on our implementation aimed for the problem of hand pose recognition.
Źródło:
Computer Science; 2017, 18 (4); 341-356
1508-2806
2300-7036
Pojawia się w:
Computer Science
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Track finding with Deep Neural Networks
Autorzy:
Kucharczyk, Marcin
Wolter, Marcin
Powiązania:
https://bibliotekanauki.pl/articles/305791.pdf
Data publikacji:
2019
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Tematy:
deep neural networks
machine learning
tracking
HEP
Opis:
High energy physics experiments require fast and efficient methods for reconstructing the tracks of charged particles. The commonly used algorithms are sequential and the required CPU power increases rapidly with the number of tracks. Neural networks can speed up the process due to their capability of modeling complex non-linear data dependencies and finding all tracks in parallel. In this paper, we describe the application of the deep neural network for reconstructing straight tracks in a toy two-dimensional model. It is planned to apply this method to the experimental data obtained by the MUonE experiment at CERN.
Źródło:
Computer Science; 2019, 20 (4); 475-491
1508-2806
2300-7036
Pojawia się w:
Computer Science
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Automatyczne rozpoznawanie treści nielegalnych filmów typu CSAM za pomocą klasyfikatora częściowo splatającego kolejne klatki materiału wideo
Autorzy:
Laskowska, Barbara
Powiązania:
https://bibliotekanauki.pl/articles/20311657.pdf
Data publikacji:
2023-10-31
Wydawca:
Akademia Sztuki Wojennej
Tematy:
cybersecurity
computer video analysis
machine learning
neural networks
deep neural networks
transfer learning
Child Sexual Abuse Material
CSAM
Opis:
The paper describes one of the methods of automatic recognition of CSAM materials, which was tested during the research under the APAKT project. The proposed solution is based on Temporal Shift Module (TSM), a model of a deep neural network created for efficient human activities rocognition in video. We applied transfer learning method for training the model with a relatively small number of training data to succesfully rocognize films with pornografic and illegal content. We conducted some tests of classification of films from three categories: neutral films, legal pornography and illegal pornografic videos (CSAM). In this paper we present problems that are connected with this research topic that come from the characteristic of the data. We also show that further works are needed to keep children safe in cyberspace.
Źródło:
Cybersecurity and Law; 2023, 10, 2; 195-201
2658-1493
Pojawia się w:
Cybersecurity and Law
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Analiza wydajności biblioteki TensorFlow z wykorzystaniem różnych algorytmów optymalizacji
Performance analysis of the TensorFlow library with different optimisation algorithms
Autorzy:
Wadas, Maciej
Smołka, Jakub
Powiązania:
https://bibliotekanauki.pl/articles/2055131.pdf
Data publikacji:
2021
Wydawca:
Politechnika Lubelska. Instytut Informatyki
Tematy:
uczenie maszynowe
sieci neuronowe
machine learning
neural networks
Opis:
W artykule zaprezentowano wyniki analizy wydajności biblioteki TensorFlow wykorzystywanej w uczeniu maszyno-wym i głębokich sieciach neuronowych. Analiza skupia się na porównaniu parametrów otrzymanych podczas treningu modelu sieci neuronowej dla algorytmów optymalizacji: Adam, Nadam, AdaMax, AdaDelta, AdaGrad. Zwrócono szczególną uwagę na różnice pomiędzy efektywnością treningu na zadaniach wykorzystujących mikroprocesor i kartę graficzną. Do przeprowadzenia badań utworzono modele sieci neuronowej, której zadaniem było rozpoznawanie znaków języka polskiego pisanych odręcznie. Otrzymane wyniki wykazały, że najwydajniejszym algorytmem jest AdaMax, zaś podzespół komputera wykorzystywany podczas badań wpływa jedynie na czas treningu wykorzystanego modelu sieci neuronowej.
This paper presents the results of performance analysis of the Tensorflow library used in machine learning and deep neural networks. The analysis focuses on comparing the parameters obtained when training the neural network model for optimization algorithms: Adam, Nadam, AdaMax, AdaDelta, AdaGrad. Special attention has been paid to the differences between the training efficiency on tasks using microprocessor and graphics card. For the study, neural network models were created in order to recognise Polish handwritten characters. The results obtained showed that the most efficient algorithm is AdaMax, while the computer component used during the research only affects the training time of the neural network model used.
Źródło:
Journal of Computer Sciences Institute; 2021, 21; 330--335
2544-0764
Pojawia się w:
Journal of Computer Sciences Institute
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Optimization of Machine Learning Process Using Parallel Computing
Autorzy:
Grzeszczyk, Michał K.
Powiązania:
https://bibliotekanauki.pl/articles/102525.pdf
Data publikacji:
2018
Wydawca:
Stowarzyszenie Inżynierów i Techników Mechaników Polskich
Tematy:
parallel computing
machine learning
perceptron
neural networks
OpenMP
Opis:
The aim of this paper is to discuss the use of parallel computing in the supervised machine learning processes in order to reduce the computation time. This way of computing has gained popularity because sequential computing is often insufficient for large scale problems like complex simulations or real time tasks. After presenting the foundations of machine learning and neural network algorithms as well as three types of parallel models, the author briefly characterized the development of the experiments carried out and the results obtained. The experiments on image recognition, ran on five sets of empirical data, prove a significant reduction in calculation time compared to classical algorithms. At the end, possible directions of further research concerning parallel optimization of calculation time in the supervised perceptron learning processes were shortly outlined.
Źródło:
Advances in Science and Technology. Research Journal; 2018, 12, 4; 81-87
2299-8624
Pojawia się w:
Advances in Science and Technology. Research Journal
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Recognition of sports exercises using inertial sensor technology
Autorzy:
Krutz, Pascal
Rehm, Matthias
Schlegel, Holger
Dix, Martin
Powiązania:
https://bibliotekanauki.pl/articles/30148258.pdf
Data publikacji:
2023
Wydawca:
Polskie Towarzystwo Promocji Wiedzy
Tematy:
human activity recognition
machine learning
neural networks
classifier
Opis:
Supervised learning as a sub-discipline of machine learning enables the recognition of correlations between input variables (features) and associated outputs (classes) and the application of these to previously unknown data sets. In addition to typical areas of application such as speech and image recognition, fields of applications are also being developed in the sports and fitness sector. The purpose of this work is to implement a workflow for the automated recognition of sports exercises in the Matlab® programming environment and to carry out a comparison of different model structures. First, the acquisition of the sensor signals provided in the local network and their processing is implemented. Realised functionalities include the interpolation of lossy time series, the labelling of the activity intervals performed and, in part, the generation of sliding windows with statistical parameters. The preprocessed data are used for the training of classifiers and artificial neural networks (ANN). These are iteratively optimised in their corresponding hyper parameters for the data structure to be learned. The most reliable models are finally trained with an increased data set, validated and compared with regard to the achieved performance. In addition to the usual evaluation metrics such as F1 score and accuracy, the temporal behaviour of the assignments is also displayed graphically, allowing statements to be made about potential causes of incorrect assignments. In this context, especially the transition areas between the classes are detected as erroneous assignments as well as exercises with insufficient or clearly deviating execution. The best overall accuracy achieved with ANN and the increased dataset was 93.7 %.
Źródło:
Applied Computer Science; 2023, 19, 1; 152-163
1895-3735
2353-6977
Pojawia się w:
Applied Computer Science
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
An Ensemble of Statistical Metadata and CNN Classification of Class Imbalanced Skin Lesion Data
Autorzy:
Nayak, Sachin
Vincent, Shweta
Sumathi, K.
Kumar, Om Prakash
Pathan, Sameena
Powiązania:
https://bibliotekanauki.pl/articles/2055258.pdf
Data publikacji:
2022
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Tematy:
classification
Convolutional Neural Networks
Ensemble Learning
machine learning
metadata
Opis:
Skin Cancer is one of the most widely present forms of cancer. The correct classification of skin lesions as malignant or benign is a complex process that has to be undertaken by experienced specialists. Another major issue of the class imbalance of data causes a bias in the results of classification. This article presents a novel approach to the usage of metadata of skin lesions images to classify them. The usage of techniques addresses the problem of class imbalance to nullify the imbalances. Further, the use of a convolutional neural network (CNN) is proposed to finetune the skin lesion data classification. Ultimately, it is proven that an ensemble of statistical metadata analysis and CNN usage would result in the highest accuracy of skin color classification instead of using the two techniques separately.
Źródło:
International Journal of Electronics and Telecommunications; 2022, 68, 2; 251--257
2300-1933
Pojawia się w:
International Journal of Electronics and Telecommunications
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Modelling cross-sectional tabular data using convolutional neural networks: Prediction of corporate bankruptcy in Poland
Autorzy:
Dzik-Walczak, Aneta
Odziemczyk, Maciej
Powiązania:
https://bibliotekanauki.pl/articles/1965119.pdf
Data publikacji:
2021-11-27
Wydawca:
Uniwersytet Warszawski. Wydział Nauk Ekonomicznych
Tematy:
convolutional neural networks
machine learning
simulation
bankruptcy prediction
financial indicators
Opis:
The paper deals with the topic of modelling the probability of bankruptcy of Polish enterprises using convolutional neural networks. Convolutional networks take images as input, so it was thus necessary to apply the method of converting the observation vector to a matrix. Benchmarks for convolutional networks were logit models, random forests, XGBoost, and dense neural networks. Hyperparameters and model architecture were selected based on a random search and analysis of learning curves and experiments in folded, stratified cross-validation. In addition, the sensitivity of the results to data preprocessing was investigated. It was found that convolutional neural networks can be used to analyze cross-sectional tabular data, especially for the problem of modelling the probability of corporate bankruptcy. In order to achieve good results with models based on parameters updated by a gradient (neural networks and logit), it is necessary to use appropriate preprocessing techniques. Models based on decision trees have been shown to be insensitive to the data transformations used.
Źródło:
Central European Economic Journal; 2021, 8, 55; 352-377
2543-6821
Pojawia się w:
Central European Economic Journal
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Application of artificial neural networks in predicting voter turnout based on the analysis of demographic data
Autorzy:
Michalak, Piotr
Powiązania:
https://bibliotekanauki.pl/articles/92572.pdf
Data publikacji:
2019
Wydawca:
Oddział Kartograficzny Polskiego Towarzystwa Geograficznego
Tematy:
artificial neural networks
voter turnout
machine learning
cartographic research method
Opis:
The author presents the results of research on the use of artificial neural networks in predicting voter turnout. He describes the principles of operation of artificial neural networks, as well as detailed results of two machine learning methods used to predict voter turnout. The research resulted in creation of a functional model that allows for prediction of voter turnout results with a considerable degree of accuracy. The entire research process was carried out using the cartographic research method.
Źródło:
Polish Cartographical Review; 2019, 51, 3; 109-116
2450-6974
Pojawia się w:
Polish Cartographical Review
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Ensemble machine learning methods to predict the balancing of ayurvedic constituents in the human body
Autorzy:
Rajasekar, Vani
Krishnamoorthi, Sathya
Saracevic, Muzafer
Pepic, Dzenis
Zajmovic, Mahir
Zogic, Haris
Powiązania:
https://bibliotekanauki.pl/articles/27312840.pdf
Data publikacji:
2022
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Tematy:
machine learning
artificial neural networks
diagnose
Ayurveda constituent
support vector machine
Opis:
In this paper, we demonstrate the result of certain machine-learning methods like support vector machine (SVM), naive Bayes (NB), decision tree (DT), k-nearest neighbor (KNN), artificial neural network (ANN), and AdaBoost algorithms for various performance characteristics to predict human body constituencies. Ayurveda-dosha studies have been used for a long time, but the quantitative reliability measurement of these diagnostic methods still lags. The careful and appropriate analysis leads to an effective treatment to predict human body constituencies. From an observation of the results, it is shown that the AdaBoost algorithm with hyperparameter tuning provides enhanced accuracy and recall (0.97), precision and F-score (0.96), and lower RSME values (0.64). The experimental results reveal that the improved model (which is based on ensemble-learning methods) significantly outperforms traditional methods. According to the findings, advancements in the proposed algorithms could give machine learning a promising future.
Źródło:
Computer Science; 2022, 23 (1); 117--132
1508-2806
2300-7036
Pojawia się w:
Computer Science
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
MIMO Beam Selection in 5G Using Neural Networks
Autorzy:
Ruseckas, Julius
Molis, Gediminas
Bogucka, Hanna
Powiązania:
https://bibliotekanauki.pl/articles/2055220.pdf
Data publikacji:
2021
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Tematy:
5G
context information
MIMO beam orientation
machine learning
neural networks
Opis:
In this paper, we consider cell-discovery problem in 5G millimeter-wave (mmWave) communication systems using multiple input, multiple output (MIMO) beam-forming technique. Specifically, we aim at the proper beam selection method using context-awareness of the user-equipment to reduce latency in beam/cell identification. Due to high path-loss in mmWave systems, beam-forming technique is extensively used to increase Signal-to-Noise Ratio (SNR). When seeking to increase user discovery distance, narrow beam must be formed. Thus, a number of possible beam orientations and consequently time needed for the discovery increases significantly when random scanning approach is used. The idea presented here is to reduce latency by employing artificial intelligence (AI) or machine learning (ML) algorithms to guess the best beam orientation using context information from the Global Navigation Satellite System (GNSS), lidars and cameras, and use the knowledge to swiftly initiate communication with the base station. To this end, here, we propose a simple neural network to predict beam orientation from GNSS and lidar data. Results show that using only GNSS data one can get acceptable performance for practical applications. This finding can be useful for user devices with limited processing power.
Źródło:
International Journal of Electronics and Telecommunications; 2021, 67, 4; 693--698
2300-1933
Pojawia się w:
International Journal of Electronics and Telecommunications
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Application of long short term memory neural networks for GPS satellite clock bias prediction
Autorzy:
Gnyś, Piotr
Przestrzelski, Paweł
Powiązania:
https://bibliotekanauki.pl/articles/1987078.pdf
Data publikacji:
2021-12-30
Wydawca:
Politechnika Gdańska
Tematy:
neural networks
LSTM
time series prediction
clock bias
GNSS
machine learning
Opis:
Satellite-based localization systems like GPS or Galileo are one of the most commonly used tools in outdoor navigation. While for most applications, like car navigation or hiking, the level of precision provided by commercial solutions is satisfactory it is not always the case for mobile robots. In the case of long-time autonomy and robots that operate in remote areas battery usage and access to synchronization data becomes a problem. In this paper, a solution providing a real-time onboard clock synchronization is presented. Results achieved are better than the current state-of-the-art solution in real-time clock bias prediction for most satellites.
Źródło:
TASK Quarterly. Scientific Bulletin of Academic Computer Centre in Gdansk; 2021, 25, 4; 381-395
1428-6394
Pojawia się w:
TASK Quarterly. Scientific Bulletin of Academic Computer Centre in Gdansk
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Machine learning and artificial intelligence techniques for detecting driver drowsiness
Autorzy:
Prathap, Boppuru Rudra
Kumar, Kukatlapalli Pradeep
Hussain, Javid
Chowdary, Cherukuri Ravindranath
Powiązania:
https://bibliotekanauki.pl/articles/27314194.pdf
Data publikacji:
2022
Wydawca:
Sieć Badawcza Łukasiewicz - Przemysłowy Instytut Automatyki i Pomiarów
Tematy:
artificial intelligence
machine learning
drowsiness detection
image processing
convolutional neural networks
AI visuals
Opis:
The number of automobiles on the road grows in lockstep with the advancement of vehicle manufacturing. Road accidents appear to be on the rise, owing to this growing proliferation of vehicles. Accidents frequently occur in our daily lives, and are the top ten causes of mortality from injuries globally. It is now an important component of the worldwide public health burden. Every year, an estimated 1.2 million people are killed in car accidents. Driver drowsiness and weariness are major contributors to traffic accidents this study relies on computer software and photographs, as well as a Convolutional Neural Network (CNN), to assess whether a motorist is tired. The Driver Drowsiness System is built on the MultiLayer Feed-Forward Network concept CNN was created using around 7,000 photos of eyes in both sleepiness and non-drowsiness phases with various face layouts. These photos were divided into two datasets: training (80% of the images) and testing (20% of the images). For training purposes, the pictures in the training dataset are fed into the network. To decrease information loss as much as feasible, backpropagation techniques and optimizers are applied. We developed an algorithm to calculate ROI as well as track and evaluate motor and visual impacts.
Źródło:
Journal of Automation Mobile Robotics and Intelligent Systems; 2022, 16, 2; 64--73
1897-8649
2080-2145
Pojawia się w:
Journal of Automation Mobile Robotics and Intelligent Systems
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Indoor localization based on visible light communication and machine learning algorithms
Autorzy:
Ghonim, Alzahraa M.
Salama, Wessam M.
Khalaf, Ashraf A. M.
Shalaby, Hossam M. H.
Powiązania:
https://bibliotekanauki.pl/articles/2063908.pdf
Data publikacji:
2022
Wydawca:
Polska Akademia Nauk. Stowarzyszenie Elektryków Polskich
Tematy:
free-space optical communication
visible light communication
neural networks
random forests
machine learning
Opis:
An indoor localization system is proposed based on visible light communications, received signal strength, and machine learning algorithms. To acquire an accurate localization system, first, a dataset is collected. The dataset is then used with various machine learning algorithms for training purpose. Several evaluation metrics are used to estimate the robustness of the proposed system. Specifically, authors’ evaluation parameters are based on training time, testing time, classification accuracy, area under curve, F1-score, precision, recall, logloss, and specificity. It turned out that the proposed system is featured with high accuracy. The authors are able to achieve 99.5% for area under curve, 99.4% for classification accuracy, precision, F1, and recall. The logloss and precision are 4% and 99.7%, respectively. Moreover, root mean square error is used as an additional performance evaluation averaged to 0.136 cm.
Źródło:
Opto-Electronics Review; 2022, 30, 2; art. no. e140858
1230-3402
Pojawia się w:
Opto-Electronics Review
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Convolutional Neural Networks for C. Elegans Muscle Age Classification Using Only Self-learned Features
Autorzy:
Czaplewski, Bartosz
Dzwonkowski, Mariusz
Panas, Damian
Powiązania:
https://bibliotekanauki.pl/articles/2176172.pdf
Data publikacji:
2022
Wydawca:
Instytut Łączności - Państwowy Instytut Badawczy
Tematy:
biomedical imaging
C. elegans muscle aging
convolutional neural networks
deep learning
machine learning
Opis:
Nematodes Caenorhabditis elegans (C. elegans) have been used as model organisms in a wide variety of biological studies, especially those intended to obtain a better understanding of aging and age-associated diseases. This paper focuses on automating the analysis of C. elegans imagery to classify the muscle age of nematodes based on the known and well established IICBU dataset. Unlike many modern classification methods, the proposed approach relies on deep learning techniques, specifically on convolutional neural networks (CNNs), to solve the problem and achieve high classification accuracy by focusing on non-handcrafted self-learned features. Various networks known from the ImageNet Large Scale Visual Recognition Challenge (ILSVRC) have been investigated and adapted for the purposes of the C. elegans muscle aging dataset by applying transfer learning and data augmentation techniques. The proposed approach of unfreezing different numbers of convolutional layers at the feature extraction stage and introducing different structures of newly trained fully connected layers at the classification stage, enable to better fine-tune the selected networks. The adjusted CNNs, as featured in this paper, have been compared with other state-of-art methods. In anti-aging drug research, the proposed CNNs would serve as a very fast and effective age determination method, thus leading to reductions in time and costs of laboratory research.
Źródło:
Journal of Telecommunications and Information Technology; 2022, 4; 85--94
1509-4553
1899-8852
Pojawia się w:
Journal of Telecommunications and Information Technology
Dostawca treści:
Biblioteka Nauki
Artykuł
Wyświetlanie 1-15 z 37

Ta witryna wykorzystuje pliki cookies do przechowywania informacji na Twoim komputerze. Pliki cookies stosujemy w celu świadczenia usług na najwyższym poziomie, w tym w sposób dostosowany do indywidualnych potrzeb. Korzystanie z witryny bez zmiany ustawień dotyczących cookies oznacza, że będą one zamieszczane w Twoim komputerze. W każdym momencie możesz dokonać zmiany ustawień dotyczących cookies