Informacja

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

Wyszukujesz frazę "Deep learning" wg kryterium: Wszystkie pola


Tytuł:
Metody deep-learning w rozpoznawaniu mowy
Melody deep-learning w rozpoznawaniu mowy
Autorzy:
Kowenzowski, Piotr
Opis:
In the last five years “Machine Learning”, “Big Data”, “Neural Nets”, etc. become buzz wordsand hot topics in science and industry. Business often doesn’t understand them. Nevertheless,a great amount of money is put in the area of “Big Data”. Development of the new algorithms andtechniques is highly desirable, because some problems (e.g. NP problems) are far too complex forprogrammers to calculate all possible combinations (like k-means clustering published by Lloydin 1957) or simple direct method does not exist (part-of-speech tagging).Machine learning (ML) scientific term was coined by Samuel (1959) as “Programmingcomputers to learn from experience should eventually eliminate the need for much of this detailedprogramming effort”. Deep learning (DL) one of the most prominent and recent ML techniqueshas brought the improvement of results and it is widely used by the industry.In this work we will adopt DL to Automatic speech recognition (ASR) in thiswork and show its advantageous over most common technique which is the Gaussianmixture models (GMM).This work is divided into three parts:Automatic speech recognition defines basic concepts and is briefly describes historyof a speech recognition. One can find the idea which will be proved later. Also it contains asummary of approaches which has been used so far but on the smaller scale.Models and features extraction shall give definitions of all algorithms. This is thebiggest part of this thesis.Experiment, data and results that gives origin of the data, descriptions of all allmetrics used in experiments, summary of our experiments and final discussion which will allowto confirm or reject our thesis.
W ciągu ostatnich lat "Uczenie Maszynowe", "Big Data", "Sieci Neuronowe" itd. stały się modnymi tematami w nauce i biznesie. Nie mniej jednak, dużo pieniędzy jest wkładanych w "Big Data", mimo że ten temat nie jest do końca rozumiany przez ludzi nie rozumiejących tego tematu. Rozwój nowych algorytmów i technik jest wysoce porządany, ponieważ niektóre problemy (np. problemy przetwarzanie języka naturalnego) są zbyt skomplikowane dla programisty, żeby policzyć wszystkie możliwe kombinacje.Głębokie uczenie (ang. Deep Learning - DL) stało się obiecującą techniką, która przyniosła przełom w rozpoznawaniu mowy. W tej pracy użyjemy technik głębokiego uczenia do rozpoznawania mowy i pokażemy jego przewagę nad dotychczas stowanymi technikami.Ta praca jest podzielona na trzy częsci:Automatyczne rozpoznawanie mowy definiuje podstawowe idee i krótko opisuje historię rozpoznawania mowy.Modele i ekstracja cech wprowadza podstawowe definicje wszystkich użytych algorytmów.Eksperymenty, dane i rezultaty opowiada o pochodzeniu danych, opisuje wszystkie metryki użyte w eksperymentach, podsumowuje eksperymenty i obala lub potwierdza tezę.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Inne
Tytuł:
Deep learning: theory and practice
Autorzy:
Cichocki, A.
Poggio, T.
Osowski, S.
Lempitsky, V.
Tematy:
deep learning
networks
theory
practice
uczenie głębokie
sieci
teoria
praktyka
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/202346.pdf  Link otwiera się w nowym oknie
Opis:
This Special Section of the Bulletin of the Polish Academy of Sciences on Technical Sciences is devoted to theoretical aspects of deep machine learning as well as practical applications in some areas of signal and image processing, particularly in bioengineering.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Student Responsibility In Deep Learning
Autorzy:
Ciechanowska, Dorota
Wydawca:
Oficyna Wydawnicza “Humanitas” Sosnowiec
Cytata wydawnicza:
D.Ciechanowska, Student Responsibility In Deep Learning, [in:] EDUCATION OF TOMORROW. Since Education in Family To System Aspects of Education,(Ed.) K.Denek, A.Kamińska, P.Oleśniewicz,
Opis:
Universities are faced with the necessity of redefining their educational goals in relation to the newly-defined tasks posed to the academic education. The outcomes of university education defined in the language of competence will not be achieved, unless the university ceases to continue the transmission strategies of teaching. Activation of students in the process of education means a shift in emphasis from teaching to make students responsible for their learning process. Autonomy in learning leads one to self-directedness and deep learning, which involves critical analysis of new information and combining it with the existing memory concepts, building the personal knowledge of the student.
Dorota Ciechanowska
Dostawca treści:
Repozytorium Centrum Otwartej Nauki
Artykuł
Tytuł:
Deep learning methods in electroencephalography
Autorzy:
Stąpor, Katarzyna
Ochab, Jeremi
Kotowski, Krzysztof
Wydawca:
Springer
Opis:
The volume, variability and high level of noise in electroencephalographic (EEG) recordings of the electrical brain activity make them difficult to approach with standard machine learning techniques. Deep learning methods, especially artificial neural networks inspired by the structure of the brain itself are better suited for the domain because of their end-to-end approach. They have already shown outstanding performance in computer vision and they are increasingly popular in the EEG domain. In this chapter, the state-of-the-art architectures and approaches to classification, segmentation, and enhancement of EEG recordings are described in applications to brain-computer interfaces, medical diagnostics and emotion recognition. In the experimental part, the complete pipeline of deep learning for EEG is presented on the example of the detection of erroneous responses in the Eriksen flanker task with results showing advantages over a traditional machine learning approach. Additionally, the refined list of public EEG data sources suitable for deep learning and guidelines for future applications are given.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Theory II: Deep learning and optimization
Autorzy:
Poggio, T.
Liao, Q.
Tematy:
deep learning
convolutional neural networks
loss surface
optimization
uczenie głębokie
sieć neuronowa
optymalizacja
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/201787.pdf  Link otwiera się w nowym oknie
Opis:
The landscape of the empirical risk of overparametrized deep convolutional neural networks (DCNNs) is characterized with a mix of theory and experiments. In part A we show the existence of a large number of global minimizers with zero empirical error (modulo inconsistent equations). The argument which relies on the use of Bezout theorem is rigorous when the RELUs are replaced by a polynomial nonlinearity. We show with simulations that the corresponding polynomial network is indistinguishable from the RELU network. According to Bezout theorem, the global minimizers are degenerate unlike the local minima which in general should be non-degenerate. Further we experimentally analyzed and visualized the landscape of empirical risk of DCNNs on CIFAR-10 dataset. Based on above theoretical and experimental observations, we propose a simple model of the landscape of empirical risk. In part B, we characterize the optimization properties of stochastic gradient descent applied to deep networks. The main claim here consists of theoretical and experimental evidence for the following property of SGD: SGD concentrates in probability – like the classical Langevin equation – on large volume, ”flat” minima, selecting with high probability degenerate minimizers which are typically global minimizers.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Skin lesion detection using deep learning
Autorzy:
Chandra, Rajit
Hajiarbabi, Mohammadreza
Tematy:
skin lesion
DenseNet
Inception V3
Pokaż więcej
Wydawca:
Sieć Badawcza Łukasiewicz - Przemysłowy Instytut Automatyki i Pomiarów
Powiązania:
https://bibliotekanauki.pl/articles/27314216.pdf  Link otwiera się w nowym oknie
Opis:
Skin lesion can be deadliest if not detected early. Early detection of skin lesion can save many lives. Artificial Intelligence and Machine learning is helping healthcare in many ways and so in the diagnosis of skin lesion. Computer aided diagnosis help clinicians in detecting the cancer. The study was conducted to classify the seven classes of skin lesion using very powerful convolutional neural networks. The two pre trained models i.e DenseNet and Incepton-v3 were employed to train the model and accuracy, precision, recall, f1score and ROCAUC was calculated for every class prediction. Moreover, gradient class activation maps were also used to aid the clinicians in determining what are the regions of image that influence model to make a certain decision. These visualizations are used for explain ability of the model. Experiments showed that DenseNet performed better then Inception V3. Also it was noted that gradient class activation maps highlighted different regions for predicting same class. The main contribution was to introduce medical aided visualizations in lesion classification model that will help clinicians in understanding the decisions of the model. It will enhance the reliability of the model. Also, different optimizers were employed with both models to compare the accuracies.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Prediction and classification of pressure injuries by deep learning
Wykrywanie i klasyfikacja odleżyn z wykorzystaniem deep learning
Autorzy:
Yilmaz, A.
Kızıl, H.
Kaya, U.
Cakır, R.
Demiral, M.
Tematy:
deep learning
pressure ulcers
artificial intelligence
nursing care
odleżyny
sztuczna inteligencja
opieka pielęgniarska
Pokaż więcej
Wydawca:
Akademia Bialska Nauk Stosowanych im. Jana Pawła II w Białej Podlaskiej
Powiązania:
https://bibliotekanauki.pl/articles/2047948.pdf  Link otwiera się w nowym oknie
Opis:
Pressure injuries are a serious medical problem that both negatively affects the patient's quality of life and results in significant healthcare costs. In cases where a patient doesn’t receive appropriate treatment and care, death may result. Nurses play critical roles in the prevention, care, and treatment of pressure injuries as members of the healthcare team who closely monitor the health status of the patient. Today, the use of artificial intelligence is becoming more prevalent in healthcare, as in many other areas. Artificial intelligence is a method that aims to solve complex problems by using computers to mathematically simulate the way the brain works. In this article, we compile and share information about a deep learning model developed for the detection and classification of pressure injuries. Deep learning can operate on many types of data. Convolutional Neural Networks (CNN) prefer images because they can handle 2D arrays. In this case, the images, annotated according to the National Pressure Injury Advisory Panel pressure injury classification system, have been fed into a deep learning model using CNN. The developed CNN model has a 97% success in detecting and classifying pressure injuries, and as more images are collected and fed into the CNN, the prediction accuracy will increase. This deep learning model allows for the automatic detection and classification of pressure injuries, an indicator of health outcomes, at an early stage and for quick and accurate intervention. In this context, it is expected that the quality of nursing care will increase, the prevalence of pressure injury will decrease, and the economic burden of this health problem will decrease.
Odleżyny są problemem zdrowotnym, który negatywnie wpływa na jakość życia pacjenta i powoduje poważne koszty opieki. W przypadku braku odpowiedniego leczenia i opieki może to doprowadzić do śmierci pacjenta. Pielęgniarki odgrywają kluczową rolę w zapobieganiu, opiece i leczeniu odleżyn jako członkowie zespołu opieki zdrowotnej, którzy ściśle i stale monitorują stan zdrowia danej osoby. Obecnie w dziedzinie zdrowia, podobnie jak w wielu innych dziedzinach, coraz częściej wykorzystuje się sztuczną inteligencję. Sztuczna inteligencja jest metodą, która ma na celu rozwiązywanie złożonych problemów poprzez matematyczne symulowanie sposobu działania mózgu z wykorzystaniem komputerów. Niniejszy artykuł jest przeglądem zaprojektowanym w celu podzielenia się informacjami na temat modelu deep learning opracowanego do wykrywania i klasyfikacji odleżyn. Deep learning może działać na wielu typach danych. Konwolucyjne sieci neuronowe (ang. convolutional neural networks, CNN) preferują obrazy, ponieważ mogą obsługiwać macierze 2D. Obrazy, uporządkowane zgodnie z systemem klasyfikacji odleżyn według National Pressure Injury Advisory Panel (NPIAP), zostały przekształcone w "Deep Learning Model" z wykorzystaniem CNN. Opracowywany model CNN ma 97% skuteczności w wykrywaniu i klasyfikowaniu odleżyn, a im więcej obrazów zostanie zebranych i wykorzystanych w CNN, tym większe będzie prawdopodobieństwo trafnej prognozy. Ten model deep learning daje możliwość automatycznego wykrywania i klasyfikacji odleżyn, które są wskaźnikiem jakości zdrowia, na wczesnym etapie oraz dokładnej i szybkiej interwencji. W tym kontekście oczekuje się, że jakość opieki pielęgniarskiej wzrośnie, zmniejszy się częstość występowania odleżyn oraz obciążenie ekonomiczne związane z tym problemem zdrowotnym.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Multi agent deep learning with cooperative communication
Autorzy:
Simões, David
Lau, Nuno
Reis, Luís Paulo
Tematy:
multi-agent systems
deep reinforcement learning
centralized learning
Pokaż więcej
Wydawca:
Społeczna Akademia Nauk w Łodzi. Polskie Towarzystwo Sieci Neuronowych
Powiązania:
https://bibliotekanauki.pl/articles/1837537.pdf  Link otwiera się w nowym oknie
Opis:
We consider the problem of multi agents cooperating in a partially-observable environment. Agents must learn to coordinate and share relevant information to solve the tasks successfully. This article describes Asynchronous Advantage Actor-Critic with Communication (A3C2), an end-to-end differentiable approach where agents learn policies and communication protocols simultaneously. A3C2 uses a centralized learning, distributed execution paradigm, supports independent agents, dynamic team sizes, partiallyobservable environments, and noisy communications. We compare and show that A3C2 outperforms other state-of-the-art proposals in multiple environments.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning in pharmacology: opportunities and threats
Autorzy:
Kocić, Ivan
Kocić, Milan
Rusiecka, Izabela
Kocić, Adam
Kocić, Eliza
Tematy:
machine learning
pharmacology
deep learning
artificial intelligence
drug research and development
Pokaż więcej
Wydawca:
Gdański Uniwersytet Medyczny
Powiązania:
https://bibliotekanauki.pl/articles/25728738.pdf  Link otwiera się w nowym oknie
Opis:
Introduction This review aims to present briefly the new horizon opened to pharmacology by the deep learning (DL) technology, but also to underline the most important threats and limitations of this method. Material and Methods We searched multiple databases for articles published before May 2021 according to the preferred reported item related to deep learning and drug research. Out of the 267 articles retrieved, we included 50 in the final review. Results DL and other different types of artificial intelligence have recently entered all spheres of science, taking an increasingly central position in the decision-making processes, also in pharmacology. Hence, there is a need for better understanding of these technologies. The basic differences between AI (artificial intelligence), DL and ML (machine learning) are explained. Additionally, the authors try to highlight the role of deep learning methods in drug research and development as well as in improving the safety of pharmacotherapy. Finally, future directions of DL in pharmacology were outlined as well as possible misuses of it. Conclusions DL is a promising and powerful tool for comprehensive analysis of big data related to all fields of pharmacology, however it has to be used carefully.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning-based initialization for object packing
Autorzy:
Wołczyk, Maciej
Opis:
One of the most important optimization tasks in the industry at the current time is the object packing problem. Although several methods have been developed for the purpose of solving it, they are usually only able to optimize placement locally and as such are heavily dependent on the choice of the initial setting -- hence the need for trying out multiple possible starting points, which impacts algorithm running time. In this paper we present a neural network-based model which provides sensible starting points in a linear time.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Deep learning in pharmacology: opportunities and threats
Autorzy:
Kocić, Adam
Kocić, Eliza
Kocić, Milan
Kocić, Ivan
Rusiecka, Izabela
Wydawca:
Gdański Uniwersytet Medyczny
Cytata wydawnicza:
Kocić I, Kocić M, Rusiecka I, Kocić A, Kocić E. Deep learning in pharmacology: chance and threats. Eur J Transl Clin Med. 2022;5(2):88-94. DOI: 10.31373/ejtcm/149217
Opis:
Introduction: This review aims to present briefly the new horizon opened to pharmacology by the deep learning (DL) technology, but also to underline the most important threats and limitations of this method. Material and Methods: We searched multiple databases for articles published before May 2021 according to the preferred reported item related to deep learning and drug research. Out of the 267 articles retrieved, we included 49 in the final review. Results: DL and other different types of artificial intelligence have recently entered all spheres of science, taking an increasingly central position in the decision-making processes, also in pharmacology. Hence, there is a need for better understanding of these technologies. The basic differences between AI (artificial intelligence), DL and ML (machine learning) are explained. Additionally, the authors try to highlight the role of deep learning methods in drug research and development as well as in improving the safety of pharmacotherapy. Finally, future directions of DL in pharmacology were outlined as well as possible misuses of it. Conclusion: DL is a promising and powerful tool for comprehensive analysis of big data related to all fields of pharmacology, however it has to be used carefully.
Dostawca treści:
Repozytorium Centrum Otwartej Nauki
Artykuł
Tytuł:
Automatic mechanical diagnostics using deep learning methods
Autorzy:
Dardzińska-Głębocka, Agnieszka
Kasperczuk, Anna
Dąbrowski, Jakub
Tematy:
Automatic diagnosis
Deep learning
Chest X-ray
Disease classification
Pokaż więcej
Wydawca:
Politechnika Białostocka. Oficyna Wydawnicza Politechniki Białostockiej
Powiązania:
https://bibliotekanauki.pl/chapters/58969415.pdf  Link otwiera się w nowym oknie
Opis:
The main goal of this work is to create a classification model using deep learning methods, to classify healthy, COVID-19 and viral influenza patients based on chest X-ray images. Two models were created, one used manual feature extraction based on texture features and the other performed feature extraction automatically in one of the network layers. Both models used an eight-layer artificial neural network for classification. The obtained models were then compared with each other. A window-based classification application was created based on the model that achieved higher accuracy. Deep learning methods used to create disease classifiers can significantly speed up and facilitate the diagnostic process.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
DNA recognition using Novel Deep Learning Model
Autorzy:
Al-Kaltakchi, Musab T. S.
Abdulla, Hasan A.
Al-Nima, Raid Rafi Omar
Tematy:
DNA recognition
deep learning
DNA identification
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czasopisma i Monografie PAN
Powiązania:
https://bibliotekanauki.pl/articles/58973147.pdf  Link otwiera się w nowym oknie
Opis:
DNA, a significant physiological biometric, is present in all human cells like hair, blood, and skin. This research introduces a new approach called the Deep DNA Learning Network (DDLN) for person identification based on their DNA. This novel Machine Learning model is designed to gather DNA chromosomes from an individual’s parents. The model’s flexibility allows it to expand or contract and has the capability to determine one or both parents of an individual using the provided chromosomes. Notably, the DDLN model offers quick training in comparison to traditional deep learning methods. The study employs two real datasets from Iraq: the Real Iraqi Dataset for Kurds (RIDK) and the Real Iraqi Dataset for Arabs (RIDA). The outcomes demonstrate that the proposed DDLN model achieves an Equal Error Rate (EER) of 0 for both datasets, indicating highly accurate performance.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Microservice-Oriented Workload Prediction Using Deep Learning
Autorzy:
Ştefan, Sebastian
Niculescu, Virginia
Tematy:
microservice
web service
workload prediction
performance modeling
microservice-applications
microservice scaler
Pokaż więcej
Wydawca:
Politechnika Wrocławska. Oficyna Wydawnicza Politechniki Wrocławskiej
Powiązania:
https://bibliotekanauki.pl/articles/2060924.pdf  Link otwiera się w nowym oknie
Opis:
Background: Service oriented architectures are becoming increasingly popular due to their flexibility and scalability which makes them a good fit for cloud deployments. Aim: This research aims to study how an efficient workload prediction mechanism for a practical proactive scaler, could be provided. Such a prediction mechanism is necessary since in order to fully take advantage of on-demand resources and reduce manual tuning, an auto-scaling, preferable predictive, approach is required, which means increasing or decreasing the number of deployed services according to the incoming workloads. Method: In order to achieve the goal, a workload prediction methodology that takes into account microservice concerns is proposed. Since, this should be based on a performant model for prediction, several deep learning algorithms were chosen to be analysed against the classical approaches from the recent research. Experiments have been conducted in order to identify the most appropriate prediction model. Results: The analysis emphasises very good results obtained using the MLP (MultiLayer Perceptron) model, which are better than those obtained with classical time series approaches, with a reduction of the mean error prediction of 49%, when using as data, two Wikipedia traces for 12 days and with two different time windows: 10 and 15min. Conclusion: The tests and the comparison analysis lead to the conclusion that considering the accuracy, but also the computational overhead and the time duration for prediction, MLP model qualifies as a reliable foundation for the development of proactive microservice scaler applications.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Usage of deep learning in recent applications
Autorzy:
Dubey, A.
Tematy:
conceptual based information retrieval
ontology
semantic search
wyszukiwanie informacji oparte na pojęciach
ontologia
wyszukiwanie semantyczne
Pokaż więcej
Wydawca:
Stowarzyszenie Komputerowej Nauki o Materiałach i Inżynierii Powierzchni w Gliwicach
Powiązania:
https://bibliotekanauki.pl/articles/24200557.pdf  Link otwiera się w nowym oknie
Opis:
Purpose: Deep learning is a predominant branch in machine learning, which is inspired by the operation of the human biological brain in processing information and capturing insights. Machine learning evolved to deep learning, which helps to reduce the involvement of an expert. In machine learning, the performance depends on what the expert extracts manner features, but deep neural networks are self-capable for extracting features. Design/methodology/approach: Deep learning performs well with a large amount of data than traditional machine learning algorithms, and also deep neural networks can give better results with different kinds of unstructured data. Findings: Deep learning is an inevitable approach in real-world applications such as computer vision where information from the visual world is extracted, in the field of natural language processing involving analyzing and understanding human languages in its meaningful way, in the medical area for diagnosing and detection, in the forecasting of weather and other natural processes, in field of cybersecurity to provide a continuous functioning for computer systems and network from attack or harm, in field of navigation and so on. Practical implications: Due to these advantages, deep learning algorithms are applied to a variety of complex tasks. With the help of deep learning, the tasks that had been said as unachievable can be solved. Originality/value: This paper describes the brief study of the real-world application problems domain with deep learning solutions.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Towards a deep learning model for hadronization
Autorzy:
Nachman, Benjamin
Siódmok, Andrzej
Ju, Xiangyang
Ghosh, Aishik
Opis:
Hadronization is a complex quantum process whereby quarks and gluons become hadrons. The widely used models of hadronization in event generators are based on physically inspired phenomenological models with many free parameters. We propose an alternative approach whereby neural networks are used instead. Deep generative models are highly flexible, differentiable, and compatible with graphical processing units. We make the first step towards a data-driven machine learning-based hadronization model. In that step, we replace a component of the hadronization model within the Herwig event generator (cluster model) with HADML, a computer code implementing a generative adversarial network. We show that a HADML is capable of reproducing the kinematic properties of cluster decays. Furthermore, we integrate it into Herwig to generate entire events that can be compared with the output of the public Herwig simulator as well as with $e^{+}e^{-}$ data
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Deep learning approach to bacterial colony classification
Autorzy:
Spurek, Przemysław
Zieliński, Bartosz
Brzychczy-Włoch, Monika
Plichta, Anna
Misztal, Krzysztof
Ochońska, Dorota
Opis:
In microbiology it is diagnostically useful to recognize various genera and species of bacteria. It can be achieved using computer-aided methods, which make the recognition processes more automatic and thus significantly reduce the time necessary for the classification. Moreover, in case of diagnostic uncertainty (the misleading similarity in shape or structure of bacterial cells), such methods can minimize the risk of incorrect recognition. In this article, we apply the state of the art method for texture analysis to classify genera and species of bacteria. This method uses deep Convolutional Neural Networks to obtain image descriptors, which are then encoded and classified with Support Vector Machine or Random Forest. To evaluate this approach and to make it comparable with other approaches, we provide a new dataset of images. DIBaS dataset (Digital Image of Bacterial Species) contains 660 images with 33 different genera and species of bacteria.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Improving accuracy of detecting dangerous objects with deep learning
Poprawa skuteczności wykrycia niebezpiecznych obiektów przy użyciu technik deep learning
Autorzy:
Zacniewski, A.
Tematy:
detecting dangerous objects
deep learning
detekcja niebezpiecznych obiektów
technika deep learning
Pokaż więcej
Wydawca:
Instytut Naukowo-Wydawniczy "SPATIUM"
Powiązania:
https://bibliotekanauki.pl/articles/315763.pdf  Link otwiera się w nowym oknie
Opis:
In this article, the problem of detecting dangerous objects with deep learning is presented. Convolutional Neural Networks are created with Python language ecosystem (Theano and Keras libraries), and then trained with different number of layers and different parameters. Accuracy of detection dangerous objects for artificial Neural Network with smaller number of layers is computed and obtained result is improved with deep learning. CIFAR-10 dataset is used due to useful classes included.
W artykule przedstawiono problem detekcji niebezpiecznych obiektów przy użyciu technik deep learning. Konwolucyjne sieci neuronowe tworzone są przy pomocy bibliotek języka Python takich jak Keras i Theano, a następnie trenowane są przy różnej liczbie warstw i z różnymi parametrami. Skuteczność detekcji niebezpiecznych obiektów dla małej liczby warstw sztucznej sieci neuronowej jest obliczana, a uzyskany wynik jest ulepszany przy użyciu technik deep learning. Zbiór danych CIFAR-10 został wykorzystany w badaniach z powodu dużej użyteczności występujących w nim klas.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Towards ensuring software interoperability between deep learning frameworks
Autorzy:
Lee, Youn Kyu
Park, Seong Hee
Lim, Min Young
Lee, Soo-Hyun
Jeong, Jongwook
Tematy:
deep learning
interoperability
validation
verification
deep learning framework
model conversion
Pokaż więcej
Wydawca:
Społeczna Akademia Nauk w Łodzi. Polskie Towarzystwo Sieci Neuronowych
Powiązania:
https://bibliotekanauki.pl/articles/23944833.pdf  Link otwiera się w nowym oknie
Opis:
With the widespread of systems incorporating multiple deep learning models, ensuring interoperability between target models has become essential. However, due to the unreliable performance of existing model conversion solutions, it is still challenging to ensure interoperability between the models developed on different deep learning frameworks. In this paper, we propose a systematic method for verifying interoperability between pre- and post-conversion deep learning models based on the validation and verification approach. Our proposed method ensures interoperability by conducting a series of systematic verifications from multiple perspectives. The case study confirmed that our method successfully discovered the interoperability issues that have been reported in deep learning model conversions.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep Learning Can Improve Early Skin Cancer Detection
Autorzy:
Mohamed, Abeer
Mohamed, Wael A.
Zekry, Abdel Halim
Tematy:
technology
dermoscopic lesions
convolutional
neural network
ISIC dataset
deep learning
neural networks
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/963798.pdf  Link otwiera się w nowym oknie
Opis:
Skin cancer is the most common form of cancer affecting humans. Melanoma is the most dangerous type of skin cancer; and early diagnosis is extremely vital in curing the disease. So far, the human knowledge in this field is very limited, thus, developing a mechanism capable of identifying the disease early on can save lives, reduce intervention and cut unnecessary costs. In this paper, the researchers developed a new learning technique to classify skin lesions, with the purpose of observing and identifying the presence of melanoma. This new technique is based on a convolutional neural network solution with multiple configurations; where the researchers employed an International Skin Imaging Collaboration (ISIC) dataset. Optimal results are achieved through a convolutional neural network composed of 14 layers. This proposed system can successfully and reliably predict the correct classification of dermoscopic lesions with 97.78% accuracy.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep Learning in Military Applications: Threats and Opportunities
Autorzy:
Surma, Jerzy
Tematy:
artificial intelligence
deep learning
technology
sztuczna inteligencja
głębokie uczenie się
technologia
Pokaż więcej
Wydawca:
Centrum Rzeczoznawstwa Budowlanego Sp. z o.o.
Powiązania:
https://bibliotekanauki.pl/articles/58906176.pdf  Link otwiera się w nowym oknie
Opis:
The latest advancements in Artificial Intelligence, especially in Deep Learning technology, accelerate innovation and development in different application domains. The development of Deep Learning technology has profoundly impacted military development trends, leading to major changes in the forms and models of war. In this paper, we overview Deep Learning’s history and architecture. Then, we review related work and extensively describe Deep Learning in two primary military applications: intelligence operations and autonomous platforms. Finally, we discuss related threats, opportunities, technical and practical difficulties. The main findings are that Artificial Intelligence technology is not omnipotent and needs to be applied carefully, considering its limitations, cybersecurity threats and a strong need for human supervision in the OODA decision loop. Certain safeguard mechanisms are required at the strategic decision-making level. In this context, one of the most important aspects relates to the education, training and selection of military officer personnel.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Effectiveness of Unsupervised Training in Deep Learning Neural Networks
Autorzy:
Rusiecki, Andrzej
Kordos, Mirosław
Tematy:
neural networks
deep learning
restricted Boltzmann Machine
contrastive divergence
Pokaż więcej
Wydawca:
Uniwersytet Jagielloński. Wydawnictwo Uniwersytetu Jagiellońskiego
Powiązania:
https://bibliotekanauki.pl/articles/1373690.pdf  Link otwiera się w nowym oknie
Opis:
Deep learning is a field of research attracting nowadays much attention, mainly because deep architectures help in obtaining outstanding results on many vision, speech and natural language processing – related tasks. To make deep learning effective, very often an unsupervised pretraining phase is applied. In this article, we present experimental study evaluating usefulness of such approach, testing on several benchmarks and different percentages of labeled data, how Contrastive Divergence (CD), one of the most popular pretraining methods, influences network generalization.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning based road recognition for intelligent suspension systems
Autorzy:
Sun, Jinwei
Cong, Jingyu
Tematy:
intelligent suspension system
deep learning
road recognition
Pokaż więcej
Wydawca:
Polskie Towarzystwo Mechaniki Teoretycznej i Stosowanej
Powiązania:
https://bibliotekanauki.pl/articles/2055054.pdf  Link otwiera się w nowym oknie
Opis:
This paper presents a deep learning-based road recognition strategy for advanced suspension systems. A four-quarter suspension model with a magnetorheological (MR) damper is developed, and four typical road images with corresponding roughness data are collected. A back-propagation neural network based autoencoder and Convolutional Neural Networks (CNN) are utilized to form the deep learning structure. By utilizing the multi-object genetic algorithm, the optimal parameters can be obtained, and the control current can be adaptively adjusted. Simulation results indicate that the designed structure can identify the road type accurately, and the recognition-based control strategy can improve the suspension performance effectively.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning based Tamil Parts of Speech (POS) tagger
Autorzy:
Anbukkarasi, S.
Varadhaganapathy, S.
Tematy:
POS tagging
part of speech
deep learning
natural language processing
BiLSTM
Bi-directional long short term memory
tagowanie POS
części mowy
uczenie głębokie
przetwarzanie języka naturalnego
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2086879.pdf  Link otwiera się w nowym oknie
Opis:
This paper addresses the problem of part of speech (POS) tagging for the Tamil language, which is low resourced and agglutinative. POS tagging is the process of assigning syntactic categories for the words in a sentence. This is the preliminary step for many of the Natural Language Processing (NLP) tasks. For this work, various sequential deep learning models such as recurrent neural network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU) and Bi-directional Long Short-Term Memory (Bi-LSTM) were used at the word level. For evaluating the model, the performance metrics such as precision, recall, F1-score and accuracy were used. Further, a tag set of 32 tags and 225 000 tagged Tamil words was utilized for training. To find the appropriate hidden state, the hidden states were varied as 4, 16, 32 and 64, and the models were trained. The experiments indicated that the increase in hidden state improves the performance of the model. Among all the combinations, Bi-LSTM with 64 hidden states displayed the best accuracy (94%). For Tamil POS tagging, this is the initial attempt to be carried out using a deep learning model.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Predicting stock market by sentiment analysis and deep learning
Autorzy:
Özögür Akyüz, Süreyya
Karadayı Ataş, Pınar
Benkhaldoun, Aymane
Tematy:
stock market
Twitter
deep learning
sentiment analysis
Pokaż więcej
Wydawca:
Politechnika Wrocławska. Oficyna Wydawnicza Politechniki Wrocławskiej
Powiązania:
https://bibliotekanauki.pl/articles/58969809.pdf  Link otwiera się w nowym oknie
Opis:
The stock market may be unpredictable; understanding when to purchase and sell can greatly assist businesses and individuals in maximizing profits and minimizing losses. Many companies have previously modified time-series analysis, a data mining technique, to forecast stock price movement. The idea of textual data mining has recently come up in debates about stock market forecasts. In this study, five of the largest firms’ historical stock prices were used to train two deep learning models—long short-term memory (LSTM) and one-dimensional convolutional neural network (1D CNN), then the results of all the models were compared. To connect price value fluctuations with the general public, sentiment scores were offered in addition to stock price values by employing natural language processing techniques (TextBlob) to tweets.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Hospitalization patient forecasting based on multi-task deep learning
Autorzy:
Zhou, Min
Huang, Xiaoxiao
Liu, Haipeng
Zheng, Dingchang
Tematy:
hospitalization patient
neural network
multitask learning
pacjent hospitalizowany
sieć neuronowa
nauka wielozadaniowa
Pokaż więcej
Wydawca:
Uniwersytet Zielonogórski. Oficyna Wydawnicza
Powiązania:
https://bibliotekanauki.pl/articles/2201025.pdf  Link otwiera się w nowym oknie
Opis:
Forecasting the number of hospitalization patients is important for hospital management. The number of hospitalization patients depends on three types of patients, namely, admission patients, discharged patients, and inpatients. However, previous works focused on one type of patients rather than the three types of patients together. In this paper, we propose a multi-task forecasting model to forecast the three types of patients simultaneously. We integrate three neural network modules into a unified model for forecasting. Besides, we extract date features of admission and discharged patient flows to improve forecasting accuracy. The algorithm is trained and evaluated on a real-world data set of a one-year daily observation of patient numbers in a hospital. We compare the performance of our model with eight baselines over two real-word data sets. The experimental results show that our approach outperforms other baseline algorithms significantly.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Hybrid deep learning method for detection of liver cancer
Autorzy:
Deshmukh, Sunita P.
Choudhari, Dharmaveer
Amalraj, Shankar
Matte, Pravin N.
Tematy:
liver cancer detection
deep learning
fully convolutional neural network
hybrid approach
discrete wavelet transform
wykrywanie raka wątroby
uczenie głębokie
neuronowa sieć konwulcyjna
podejście hybrydowe
dyskretna transformata falkowa
Pokaż więcej
Wydawca:
Instytut Podstawowych Problemów Techniki PAN
Powiązania:
https://bibliotekanauki.pl/articles/38701864.pdf  Link otwiera się w nowym oknie
Opis:
Liver disease refers to any liver irregularity causing its damage. There are several kinds of liver ailments. Benign growths are rarely life threatening and can be removed by specialists. Liver malignant tumor is leading causes of cancer death. Identifying malignant growth tissue is a troublesome and tedious task. There is significantly less information and statistical analysis presented related to cholangiocarcinoma and hepatoblastoma. This research focuses on the image analysis of these two types of cancer. The framework’s performance is evaluated using 2871 images, and a dual hybrid model is used to accomplish superb exactness. The aftereffects of both neural networks are sent into the result prioritizer that decides the most ideal choice for image arrangement. The relevance of elements appears to address the appropriate imaging rules for each class, and feature maps matching the original picture voxel features. The significance of features represents the most important imaging criteria for each class. This deep learning system demonstrates the concept of illuminating elements of a pre-trained deep neural network’s decision-making process by an examination of inner layers and the description of attributes that contribute to predictions.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning versus classical neural approach to mammogram recognition
Autorzy:
Kurek, J.
Świderski, B.
Osowski, S.
Kruk, M.
Barhoumi, W.
Tematy:
convolutional neural networks
breast cancer diagnosis
mammogram recognition
diagnostic features
splotowe sieci neuronowe
diagnostyka raka piersi
rozpoznawanie
mammografia
cechy diagnostyczne
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/200919.pdf  Link otwiera się w nowym oknie
Opis:
Automatic recognition of mammographic images in breast cancer is a complex issue due to the confusing appearance of some perfectly normal tissues which look like masses. The existing computer-aided systems suffer from non-satisfactory accuracy of cancer detection. This paper addresses this problem and proposes two alternative techniques of mammogram recognition: the application of a variety of methods for definition of numerical image descriptors in combination with an efficient SVM classifier (so-called classical approach) and application of deep learning in the form of convolutional neural networks, enhanced with additional transformations of input mammographic images. The key point of the first approach is defining the proper numerical image descriptors and selecting the set which is the most class discriminative. To achieve better performance of the classifier, many image descriptors were defined by means of applying different characterization of the images: Hilbert curve representation, Kolmogorov-Smirnov statistics, the maximum subregion principle, percolation theory, fractal texture descriptors as well as application of wavelet and wavelet packets. Thanks to them, better description of the basic image properties has been obtained. In the case of deep learning, the features are automatically extracted as part of convolutional neural network learning. To get better quality of results, additional representations of mammograms, in the form of nonnegative matrix factorization and the self-similarity principle, have been proposed. The methods applied were evaluated based on a large database composed of 10,168 regions of interest in mammographic images taken from the DDSM database. Experimental results prove the advantage of deep learning over traditional approach to image recognition. Our best average accuracy in recognizing abnormal cases (malignant plus benign versus healthy) was 85.83%, with sensitivity of 82.82%, specificity of 86.59% and AUC = 0.919. These results are among the best for this massive database.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning-based fault diagnosis for marine centrifugal fan
Autorzy:
Li, Congyue
Hu, Yihuai
Jiang, Jiawei
Yan, Guohua
Tematy:
CEEMDAN
fault diagnosis
lightweight neural network
marine centrifugal fan
Pokaż więcej
Wydawca:
Politechnika Gdańska. Wydział Inżynierii Mechanicznej i Okrętownictwa
Powiązania:
https://bibliotekanauki.pl/articles/32917700.pdf  Link otwiera się w nowym oknie
Opis:
Marine centrifugal fans usually work in harsh environments. Their vibration signals are non-linear. The traditional fault diagnosis methods of fans require much calculation and have low operating efficiency. Only shallow fault features can be extracted. As a result, the diagnosis accuracy is not high. It is difficult to realize the end-to-end fault diagnosis. Combining the Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) and lightweight neural network, a fault classification method is proposed. First, the CEEMDAN can decompose the vibration signal into several intrinsic modal functions (IMF). Then, the original signals can be transformed into 2-D images through pseudocolour coding of the IMFs. Finally, they are fed into the lightweight neural network for fault diagnosis. By embedding a convolutional block attention module (CBAM), the ability of the network to extract critical feature information is improved. The results show that the proposed method can adaptively extract the fault characteristics of a marine centrifugal fan. While the model is lightweight, the overall diagnostic accuracy can reach 99.3%. As exploratory basic research, this method can provide a reference for intelligent fault diagnosis systems on ships.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep Learning-based Beamforming Approach Incorporating Linear Antenna Arrays
Autorzy:
Bhalke, Daulappa
Paikrao, Pavan D.
Anguera, Jaume
Tematy:
adaptive beamforming
antenna arrays
convolutional neural network
Pokaż więcej
Wydawca:
Instytut Łączności - Państwowy Instytut Badawczy
Powiązania:
https://bibliotekanauki.pl/articles/58906713.pdf  Link otwiera się w nowym oknie
Opis:
This research delves into exploring machine learning and deep learning techniques relied upon in antenna design processes. First, the general concepts of machine learning and deep learning are introduced. Then, the focus shifts to various antenna applications, such as those relying on millimeter waves. The feasibility of employing antennas in this band is examined and compared with conventional methods, emphasizing the acceleration of the antenna design process, reduction in the number of simulations, and improved computational efficiency. The proposed method is a low-complexity approach which avoids the need for eigenvalue decomposition, the procedure for computing the entire matrix inversion, as well as incorporating signal and interference correlation matrices in the weight optimization process. The experimental results clearly demonstrate that the proposed method outperforms the compared beamformers by achieving a better signal-to-interference ratio.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Uczenie głębokie w diagnostyce medycznej
Deep Learning in Medical Diagnosis
Autorzy:
Antczak, K.
Tematy:
sieci neuronowe
diagnostyka medyczna
uczenie głębokie
neural networks
medical diagnosis
deep learning
Pokaż więcej
Wydawca:
Polskie Towarzystwo Symulacji Komputerowej
Powiązania:
https://bibliotekanauki.pl/articles/404011.pdf  Link otwiera się w nowym oknie
Opis:
W pracy przeanalizowano perspektywy zastosowania metod uczenia głębokiego w diagnostyce medycznej. Jedną z kluczowych cech uczenia głębokiego jest zdolność do wyodrębniania złożonych wzorców o strukturze hierarchicznej. Wzorce takie występują również w diagnostyce, jako tak zwane diamenty diagnostyczne. Zastosowanie głębokich sieci neuronowych mogłoby poprawić jakość klasyfikatorów wykrywających choroby na podstawie objawów. Dodatkowo umożliwiłoby to sterowanie czułoscią i swoistością klasyfikatorów.
In this paper we analyze perspectives of applying deep learning methods in a field of medical diagnosis. One of key features of deep learning is ability to extract complex, hierarchical patterns. Such patterns are present also in a medical diagnosis, where they are known as diagnostic diamonds. Applying deep neural networks could increase performance of medical classifiers. Moreover, it would allow to adjust sensitivity and specificity of classifiers.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning-based framework for tumour detection and semantic segmentation
Autorzy:
Kot, Estera
Krawczyk, Zuzanna
Siwek, Krzysztof
Królicki, Leszek
Czwarnowski, Piotr
Tematy:
deep learning
medical imaging
tumour detection
semantic segmentation
image fusion
technika deep learning
głęboka nauka
obrazowanie medyczne
wykrywanie guza
segmentacja semantyczna
połączenie obrazu
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2128156.pdf  Link otwiera się w nowym oknie
Opis:
For brain tumour treatment plans, the diagnoses and predictions made by medical doctors and radiologists are dependent on medical imaging. Obtaining clinically meaningful information from various imaging modalities such as computerized tomography (CT), positron emission tomography (PET) and magnetic resonance (MR) scans are the core methods in software and advanced screening utilized by radiologists. In this paper, a universal and complex framework for two parts of the dose control process – tumours detection and tumours area segmentation from medical images is introduced. The framework formed the implementation of methods to detect glioma tumour from CT and PET scans. Two deep learning pre-trained models: VGG19 and VGG19-BN were investigated and utilized to fuse CT and PET examinations results. Mask R-CNN (region-based convolutional neural network) was used for tumour detection – output of the model is bounding box coordinates for each object in the image – tumour. U-Net was used to perform semantic segmentation – segment malignant cells and tumour area. Transfer learning technique was used to increase the accuracy of models while having a limited collection of the dataset. Data augmentation methods were applied to generate and increase the number of training samples. The implemented framework can be utilized for other use-cases that combine object detection and area segmentation from grayscale and RGB images, especially to shape computer-aided diagnosis (CADx) and computer-aided detection (CADe) systems in the healthcare industry to facilitate and assist doctors and medical care providers.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Skin Lesion Analysis Toward Melanoma Detection Using Deep Learning Techniques
Autorzy:
Sherif, Fatma
Mohamed, Wael A.
Mohra, A.S.
Tematy:
melanoma
skin cancer
convolutional neural network
deep learning
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/226719.pdf  Link otwiera się w nowym oknie
Opis:
In the last few years, a great attention was paid to the deep learning Techniques used for image analysis because of their ability to use machine learning techniques to transform input data into high level presentation. For the sake of accurate diagnosis, the medical field has a steadily growing interest in such technology especially in the diagnosis of melanoma. These deep learning networks work through making coarse segmentation, conventional filters and pooling layers. However, this segmentation of the skin lesions results in image of lower resolution than the original skin image. In this paper, we present deep learning based approaches to solve the problems in skin lesion analysis using a dermoscopic image containing skin tumor. The proposed models are trained and evaluated on standard benchmark datasets from the International Skin Imaging Collaboration (ISIC) 2018 Challenge. The proposed method achieves an accuracy of 96.67% for the validation set. The experimental tests carried out on a clinical dataset show that the classification performance using deep learning-based features performs better than the state-of-the-art techniques.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning-based framework for tumour detection and semantic segmentation
Autorzy:
Kot, Estera
Krawczyk, Zuzanna
Siwek, Krzysztof
Królicki, Leszek
Czwarnowski, Piotr
Tematy:
deep learning
medical imaging
tumour detection
semantic segmentation
image fusion
technika deep learning
głęboka nauka
obrazowanie medyczne
wykrywanie guza
segmentacja semantyczna
połączenie obrazu
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2173573.pdf  Link otwiera się w nowym oknie
Opis:
For brain tumour treatment plans, the diagnoses and predictions made by medical doctors and radiologists are dependent on medical imaging. Obtaining clinically meaningful information from various imaging modalities such as computerized tomography (CT), positron emission tomography (PET) and magnetic resonance (MR) scans are the core methods in software and advanced screening utilized by radiologists. In this paper, a universal and complex framework for two parts of the dose control process – tumours detection and tumours area segmentation from medical images is introduced. The framework formed the implementation of methods to detect glioma tumour from CT and PET scans. Two deep learning pre-trained models: VGG19 and VGG19-BN were investigated and utilized to fuse CT and PET examinations results. Mask R-CNN (region-based convolutional neural network) was used for tumour detection – output of the model is bounding box coordinates for each object in the image – tumour. U-Net was used to perform semantic segmentation – segment malignant cells and tumour area. Transfer learning technique was used to increase the accuracy of models while having a limited collection of the dataset. Data augmentation methods were applied to generate and increase the number of training samples. The implemented framework can be utilized for other use-cases that combine object detection and area segmentation from grayscale and RGB images, especially to shape computer-aided diagnosis (CADx) and computer-aided detection (CADe) systems in the healthcare industry to facilitate and assist doctors and medical care providers.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
VSC-Based DSTATCOM for PQ Improvement: A Deep-Learning Approach
Autorzy:
Mangaraj, Mrutyunjaya
Sabat, Jogeswara
Barisal, Ajit Kumar
Ramaiah, K. Subba
Rao, Gudivada Eswara
Tematy:
DL approach
deep learning approach
DSTATCOM
distributed static compensator
ALMS
PQ
power quality
Pokaż więcej
Wydawca:
Politechnika Wrocławska. Oficyna Wydawnicza Politechniki Wrocławskiej
Powiązania:
https://bibliotekanauki.pl/articles/2175932.pdf  Link otwiera się w nowym oknie
Opis:
With the rapid advancement of the technology, deep learning supported voltage source converter (VSC)-based distributed static compensator (DSTATCOM) for power quality (PQ) improvement has attracted significant interest due to its high accuracy. In this paper, six subnets are structured for the proposed deep learning approach (DL-Approach) algorithm by using its own mathematical equations. Three subnets for active and the other three for reactive weight components are used to extract the fundamental component of the load current. These updated weights are utilised for the generation of the reference source currents for VSC. Hysteresis current controllers (HCCs) are employed in each phase in which generated switching signal patterns need to be carried out from both predicted reference source current and actual source current. As a result, the proposed technique achieves better dynamic performance, less computation burden and better estimation speed. Consequently, the results were obtained for different loading conditions using MATLAB/Simulink software. Finally, the feasibility was effective as per the benchmark of IEEE guidelines in response to harmonics curtailment, power factor (p.f) improvement, load balancing and voltage regulation.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Exploring Deep Learning for Underwater Plastic Debris Detection and Monitoring
Autorzy:
Khriss, Abdelaadim
Elmiad, Aissa Kerkour
Badaoui, Mohammed
Barkaoui, Alae-Eddine
Zarhloule, Yassine
Tematy:
marine debris monitoring
deep learning
YOLOv9
YOLOv8
faster rcnn
ssd
Pokaż więcej
Wydawca:
Polskie Towarzystwo Inżynierii Ekologicznej
Powiązania:
https://bibliotekanauki.pl/articles/59114363.pdf  Link otwiera się w nowym oknie
Opis:
In this paper, a comparative evaluation of state-of-the-art deep learning models for object detection in underwater environments focusing on marine debris detection was presented. The performance of four prominent object detection models was investigated, including: Faster R-CNN, SSD, YOLOv8, and YOLOv9, using two different datasets: TrashCAN and DeepTrash. Through quantitative analysis, the accuracy, precision, recall, and mean average precision (mAP) of each model across different object classes and environmental conditions were evaluated. The obtained results show that YOLOv9 consistently outperforms the other models, demonstrating superior precision, recall, and mAP values on both datasets. Furthermore, the stability and convergence behavior of the models during training were analyzed, highlighting the excellent stability and adaptability of YOLOv9. The obtained results underscore the effectiveness of deep learning-based approaches in marine debris detection and highlight the potential of YOLOv9 as a robust solution for environmental monitoring and intervention efforts in underwater ecosystems.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Research on ore fragmentation recognition method based on deep learning
Autorzy:
Jing, Hongdi
He, Wenxuan
Yu, Miao
Li, Xin
Zhang, Xingfan
Liu, Xiaosong
Cui, Yang
Wang, Zhijian
Tematy:
ruda żelaza
pękanie skały
wysadzanie skał
underground mines
ore fragmentation
visual identity
recognition
deep learning
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czasopisma i Monografie PAN
Powiązania:
https://bibliotekanauki.pl/articles/59111347.pdf  Link otwiera się w nowym oknie
Opis:
The degree of ore fragmentation in mining sites is closely related to crushing efficiency, equipment safety, beneficiation efficiency, and mining costs. Aiming to address the challenges of high labour intensity and low accuracy during manual ore fragmentation measurement at the mine site, this paper proposes a method for ore fragmentation recognition based on deep learning. This method not only uses the residual neural network structure to form the backbone feature extraction network of CSPDarkNet21 under the Darknet framework but also selects the simple two-way fusion feature PANet as the feature extraction network under the condition of only needing to identify large ore. PANet is simplified from three feature layers to one feature layer, which speeds up model training and prediction. The research results show that with a 6% decrease in accuracy, the model training time is reduced by 13 times, and the model running efficiency is improved by 21.2 times, significantly shortening the model development time. At the same time, CIOU calculates the loss value to make model training more stable. After the ore identification is completed, the real size of the ore can be obtained by calculating the pixel area of the prediction frame using the ore fragmentation judgement method.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Attention-based deep learning model for Arabic handwritten text recognition
Autorzy:
Aïcha Gader, Takwa Ben
Echi, Afef Kacem
Tematy:
Arabic handwriting recognition
attention mechanism
BLSTM
CNN
CTC
RNN
Pokaż więcej
Wydawca:
Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Instytut Informatyki Technicznej
Powiązania:
https://bibliotekanauki.pl/articles/2201264.pdf  Link otwiera się w nowym oknie
Opis:
This work proposes a segmentation-free approach to Arabic Handwritten Text Recog-nition (AHTR): an attention-based Convolutional Neural Network - Recurrent Neural Network - Con-nectionist Temporal Classification (CNN-RNN-CTC) deep learning architecture. The model receives asinput an image and provides, through a CNN, a sequence of essential features, which are transferred toan Attention-based Bidirectional Long Short-Term Memory Network (BLSTM). The BLSTM gives features sequence in order, and the attention mechanism allows the selection of relevant information from the features sequences. The selected information is then fed to the CTC, enabling the loss calculation and the transcription prediction. The contribution lies in extending the CNN by dropout layers, batch normalization, and dropout regularization parameters to prevent over-fitting. The output of the RNN block is passed through an attention mechanism to utilize the most relevant parts of the input sequence in a flexible manner. This solution enhances previous methods by improving the CNN speed and performance and controlling over model over-fitting. The proposed system achieves the best accuracy of97.1% for the IFN-ENIT Arabic script database, which competes with the current state-of-the-art. It was also tested for the modern English handwriting of the IAM database, and the Character Error Rate of 2.9% is attained, which confirms the model’s script independence.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep Learning-Powered Beamforming for 5G Massive MIMO Systems
Autorzy:
Bendjillali, Ridha Ilyas
Bendelhoum, Mohammed Sofiane
Tadjeddine, Ali Abderrazak
Kamline, Miloud
Tematy:
5G
digital beamforming
hybrid beamforming
massive MIMO
ResNeSt
Pokaż więcej
Wydawca:
Instytut Łączności - Państwowy Instytut Badawczy
Powiązania:
https://bibliotekanauki.pl/articles/27312956.pdf  Link otwiera się w nowym oknie
Opis:
In this study, a ResNeSt-based deep learning approach to beamforming for 5G massive multiple-input multipleoutput (MIMO) systems is presented. The ResNeSt-based deep learning method is harnessed to simplify and optimize the beamforming process, consequently improving performance and efficiency of 5G and beyond communication networks. A study of beamforming capabilities has revealed potential to maximize channel capacity while minimizing interference, thus eliminating inherent limitations of the traditional methods. The proposed model shows superior adaptability to dynamic channel conditions and outperforms traditional techniques across various interference scenarios.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning approach to describe and classify fungi microscopic images
Autorzy:
Rymarczyk, Dawid
Zieliński, Bartosz
Brzychczy-Włoch, Monika
Piekarczyk, Adam
Sroka-Oleksiak, Agnieszka
Opis:
Preliminary diagnosis of fungal infections can rely on microscopic examination. However, in many cases, it does not allow unambiguous identification of the species due to their visual similarity. Therefore, it is usually necessary to use additional biochemical tests. That involves additional costs and extends the identification process up to 10 days. Such a delay in the implementation of targeted therapy may be grave in consequence as the mortality rate for immunosuppressed patients is high. In this paper, we apply a machine learning approach based on deep neural networks and bag-of-words to classify microscopic images of various fungi species. Our approach makes the last stage of biochemical identification redundant, shortening the identification process by 2-3 days, and reducing the cost of the diagnosis.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Deep learning for glaucoma diagnosis
Głębokie uczenie dla diagnostyki jaskry
Autorzy:
Knapik, Andrzej
Opis:
Celem niniejszej pracy magisterskiej było stworzenie oprogramowania, które w pełni automatycznie dokonywać będzie diagnozy jaskry na podstawie kolorowych zdjęć siatkówki oka. Skupiono się na aspekcie wyznaczania z obrazu wejściowego obszaru zainteresowania zawierającego tarczę nerwu wzrokowego, a także na dalszej analizie otrzymanego fragmentu z wykorzystaniem konwolucyjnych sieci neuronowych. Stworzony klasyfikator jest w rzeczywistości złożeniem trzech sieci. Dodatkowo został zaimplementowany prosty interfejs konsolowy, umożliwiający korzystanie z aplikacji.
The purpose of this thesis was to create the software that will be able to automatically diagnose glaucoma based on retinal images. The thesis concerns the optic disc segmentation in the input image, as well as analysis of the obtained fragment with the use of convolution neural networks. The created classifier is actually example of ensemble of three networks. In addition, a simple command line interface has been implemented.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Inne
Tytuł:
Segmentation of bone structures with the use of deep learning techniques
Autorzy:
Krawczyk, Zuzanna
Starzyński, Jacek
Tematy:
deep learning
semantic segmentation
U-net
FCN
ResNet
computed tomography
technika deep learning
głęboka nauka
segmentacja semantyczna
tomografia komputerowa
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2128158.pdf  Link otwiera się w nowym oknie
Opis:
The paper is focused on automatic segmentation task of bone structures out of CT data series of pelvic region. The authors trained and compared four different models of deep neural networks (FCN, PSPNet, U-net and Segnet) to perform the segmentation task of three following classes: background, patient outline and bones. The mean and class-wise Intersection over Union (IoU), Dice coefficient and pixel accuracy measures were evaluated for each network outcome. In the initial phase all of the networks were trained for 10 epochs. The most exact segmentation results were obtained with the use of U-net model, with mean IoU value equal to 93.2%. The results where further outperformed with the U-net model modification with ResNet50 model used as the encoder, trained by 30 epochs, which obtained following result: mIoU measure – 96.92%, “bone” class IoU – 92.87%, mDice coefficient – 98.41%, mDice coefficient for “bone” – 96.31%, mAccuracy – 99.85% and Accuracy for “bone” class – 99.92%.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Segmentation of bone structures with the use of deep learning techniques
Autorzy:
Krawczyk, Zuzanna
Starzyński, Jacek
Tematy:
deep learning
semantic segmentation
U-net
FCN
ResNet
computed tomography
technika deep learning
głęboka nauka
segmentacja semantyczna
tomografia komputerowa
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2173574.pdf  Link otwiera się w nowym oknie
Opis:
The paper is focused on automatic segmentation task of bone structures out of CT data series of pelvic region. The authors trained and compared four different models of deep neural networks (FCN, PSPNet, U-net and Segnet) to perform the segmentation task of three following classes: background, patient outline and bones. The mean and class-wise Intersection over Union (IoU), Dice coefficient and pixel accuracy measures were evaluated for each network outcome. In the initial phase all of the networks were trained for 10 epochs. The most exact segmentation results were obtained with the use of U-net model, with mean IoU value equal to 93.2%. The results where further outperformed with the U-net model modification with ResNet50 model used as the encoder, trained by 30 epochs, which obtained following result: mIoU measure – 96.92%, “bone” class IoU – 92.87%, mDice coefficient – 98.41%, mDice coefficient for “bone” – 96.31%, mAccuracy – 99.85% and Accuracy for “bone” class – 99.92%.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Performance Analysis of LEACH with Deep Learning in Wireless Sensor Networks
Autorzy:
Prajapati, Hardik K.
Joshi, Rutvij
Tematy:
machine learning
Deep learning
Convolutional Neural Network (CNN)
LEACH
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2200710.pdf  Link otwiera się w nowym oknie
Opis:
Thousands of low-power micro sensors make up Wireless Sensor Networks, and its principal role is to detect and report specified events to a base station. Due to bounded battery power these nodes are having very limited memory and processing capacity. Since battery replacement or recharge in sensor nodes is nearly impossible, power consumption becomes one of the most important design considerations in WSN. So one of the most important requirements in WSN is to increase battery life and network life time. Seeing as data transmission and reception consume the most energy, it’s critical to develop a routing protocol that addresses the WSN’s major problem. When it comes to sending aggregated data to the sink, hierarchical routing is critical. This research concentrates on a cluster head election system that rotates the cluster head role among nodes with greater energy levels than the others.We used a combination of LEACH and deep learning to extend the network life of the WSN in this study. In this proposed method, cluster head selection has been performed by Convolutional Neural Network (CNN). The comparison has been done between the proposed solution and LEACH, which shows the proposed solution increases the network lifetime and throughput.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Ensemble of classifiers based on deep learning for medical image recognition
Autorzy:
Gil, Fabian
Osowski, Stanisław
Świderski, Bartosz
Słowińska, Monika
Tematy:
breast cancer
CNN
deep learning
ensemble of classifiers
feature selection
melanoma
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2203370.pdf  Link otwiera się w nowym oknie
Opis:
The paper presents special forms of an ensemble of classifiers for analysis of medical images based on application of deep learning. The study analyzes different structures of convolutional neural networks applied in the recognition of two types of medical images: dermoscopic images for melanoma and mammograms for breast cancer. Two approaches to ensemble creation are proposed. In the first approach, the images are processed by a convolutional neural network and the flattened vector of image descriptors is subjected to feature selection by applying different selection methods. As a result, different sets of a limited number of diagnostic features are generated. In the next stage, these sets of features represent input attributes for the classical classifiers: support vector machine, a random forest of decision trees, and softmax. By combining different selection methods with these classifiers an ensemble classification system is created and integrated by majority voting. In the second approach, different structures of convolutional neural networks are directly applied as the members of the ensemble. The efficiency of the proposed classification systems is investigated and compared to medical data representing dermoscopic images of melanoma and breast cancer mammogram images. Thanks to fusion of the results of many classifiers forming an ensemble, accuracy and all other quality measures have been significantly increased for both types of medical images.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A Survey on multi-objective based parameter optimization for deep learning
Autorzy:
Chakraborty, Mrittika
Pal, Wreetbhas
Bandyopadhyay, Sanghamitra
Maulik, Ujjwal
Tematy:
deep learning
multi-objective optimization
parameter optimization
neural networks
Pokaż więcej
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Powiązania:
https://bibliotekanauki.pl/articles/27312917.pdf  Link otwiera się w nowym oknie
Opis:
Deep learning models form one of the most powerful machine learning models for the extraction of important features. Most of the designs of deep neural models, i.e., the initialization of parameters, are still manually tuned. Hence, obtaining a model with high performance is exceedingly time-consuming and occasionally impossible. Optimizing the parameters of the deep networks, therefore, requires improved optimization algorithms with high convergence rates. The single objective-based optimization methods generally used are mostly time-consuming and do not guarantee optimum performance in all cases. Mathematical optimization problems containing multiple objective functions that must be optimized simultaneously fall under the category of multi-objective optimization sometimes referred to as Pareto optimization. Multi-objective optimization problems form one of the alternatives yet useful options for parameter optimization. However, this domain is a bit less explored. In this survey, we focus on exploring the effectiveness of multi-objective optimization strategies for parameter optimization in conjunction with deep neural networks. The case studies used in this study focus on how the two methods are combined to provide valuable insights into the generation of predictions and analysis in multiple applications.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Satellite Image Fusion Using a Hybrid Traditional and Deep Learning Method
Autorzy:
Hammad, Mahmoud M.
Mahmoud, Tarek A.
Amein, Ahmed Saleh
Ghoniemy, Tarek S.
Tematy:
deep learning image fusion
remote sensing image fusion
remote sensing optical image
pan-sharpening
remote sensing image
Pokaż więcej
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Powiązania:
https://bibliotekanauki.pl/articles/27314300.pdf  Link otwiera się w nowym oknie
Opis:
Due to growing demand for ground-truth in deep learning-based remote sensing satellite image fusion, numerous approaches have been presented. Of these approaches, Wald’s protocol is the most commonly used. In this paper, a new workflow is proposed consisting of two main parts. The first part targets obtaining the ground-truth images using the results of a pre-designed and well-tested hybrid traditional fusion method. This method combines the Gram–Schmidt and curvelet transform techniques to generate accurate and reliable fusion results. The second part focuses on the training of a proposed deep learning model using rich and informative data provided by the first stage to improve the fusion performance. The demonstrated deep learning model relies on a series of residual dense blocks to enhance network depth and facilitate the effective feature learning process. These blocks are designed to capture both low-level and high-level information, enabling the model to extract intricate details and meaningful features from the input data. The performance evaluation of the proposed model is carried out using seven metrics such as peak-signal-to-noise-ratio and quality without reference. The experimental results demonstrate that the proposed approach outperforms state-of-the-art methods in terms of image quality. It also exhibits the robustness and powerful nature of the proposed approach which has the potential to be applied to many remote sensing applications in agriculture, environmental monitoring, and change detection.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning method for classifying items into categories for dutch auctions
Autorzy:
Bobulski, Janusz
Szymoniak, Sabina
Tematy:
deep learning
internet auction
classification
głęboka nauka
aukcja internetowa
klasyfikacja
Pokaż więcej
Wydawca:
Instytut Podstawowych Problemów Techniki PAN
Powiązania:
https://bibliotekanauki.pl/articles/38707060.pdf  Link otwiera się w nowym oknie
Opis:
Artificial Intelligence (AI) methods are widely used in our lives (phones, social media, self-driving cars, and e-commerce). In AI methods, we can find convolutional neural networks (CNN). First of all, we can use these networks to analyze images. This paper presents a method for classifying items into particular categories on an auction site. The technique prompts the seller to which category assign the item when creating a new auction. We choose a neural network with a number of image convolution layers as the best available approach to address this task. All tests were carried out in the Matlab environment using GPU and CPU. Then, the tested and verified solution was implemented in the TensorFlow environment with a CPU processor. Thanks to the cross-validation method, the effectiveness of the recognition system was fully verified in several stages. We obtained promising results. Consequently, we implemented the developed method by adding a new sales offer on the Clemens website.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning in the classification and recognition of cardiac activity patterns
Autorzy:
Jeleń, Łukasz
Ciskowski, Piotr
Kluwak, Konrad
Tematy:
ECG signal
deep learning
arrhythmia
signal processing
ECG classification
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czasopisma i Monografie PAN
Powiązania:
https://bibliotekanauki.pl/articles/58973003.pdf  Link otwiera się w nowym oknie
Opis:
Electrocardiography is an examination performed frequently in patients experiencing symptoms of heart disease. Upon a detailed analysis, it has shown potential to detect and identify various activities. In this article, we present a deep learning approach that can be used to analyze ECG signals. Our research shows promising results in recognizing activity and disease patterns with nearly 90% accuracy. In this paper, we present the early results of our analysis, indicating the potential of using deep learning algorithms in the analysis of both onedimensional and two–dimensional data. The methodology we present can be utilized for ECG data classification and can be extended to wearable devices. Conclusions of our study pave the way for exploring live data analysis through wearable devices in order to not only predict specific cardiac conditions, but also a possibility of using them in alternative and augmented communication frameworks.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Integrated and deep learning–based social surveillance system : a novel approach
Autorzy:
Litoriya, Ratnesh
Ramchandani, Dev
Moyal, Dhruvansh
Bothra, Dhruv
Tematy:
Video Surveillance
object detection
object tracking
YOLO v4 algorithm
OpenCV
Pokaż więcej
Wydawca:
Sieć Badawcza Łukasiewicz - Przemysłowy Instytut Automatyki i Pomiarów
Powiązania:
https://bibliotekanauki.pl/articles/27314204.pdf  Link otwiera się w nowym oknie
Opis:
In industry and research, big data applications are gaining a lot of traction and space. Surveillance videos contribute significantly to big unlabelled data. The aim of visual surveillance is to understand and determine object behavior. It includes static and moving object detection, as well as video tracking to comprehend scene events. Object detection algorithms may be used to identify items in any video scene. Any video surveillance system faces a significant challenge in detecting moving objects and differentiating between objects with same shapes or features. The primary goal of this work is to provide an integrated framework for quick overview of video analysis utilizing deep learning algorithms to detect suspicious activity. In greater applications, the detection method is utilized to determine the region where items are available and the form of objects in each frame. This video analysis also aids in the attainment of security. Security may be characterized in a variety of ways, such as identifying theft or violation of covid protocols. The obtained results are encouraging and superior to existing solutions with 97% accuracy.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Recognition of font and tamil letter in images using deep learning
Autorzy:
Sridharan, Manikandan
Arulanandam, Delphin Carolina Rani
Chinnasamy, Rajeswari K
Thimmanna, Suma
Dhandapani, Sivabalaselvamani
Tematy:
deep convolution network
Tamil Letter
recognition system
font recognition
filtering
głęboka sieć konwolucyjna
system rozpoznawania
rozpoznawanie czcionek
filtrowanie
Pokaż więcej
Wydawca:
Polskie Towarzystwo Promocji Wiedzy
Powiązania:
https://bibliotekanauki.pl/articles/1837779.pdf  Link otwiera się w nowym oknie
Opis:
This paper proposes a deep learning approach to recognize Tamil Letter from images which contains text. This is recognition process, the text in the images are divided to letter or characters. Each recognized letters are sending to recognition system and filter the text using deep learning algorithms. Our proposed algorithm is used to separate letter from the text using convolution neural network approach. The filtering system is used for identifying font based on that letters are found. The Tamil letters are test data and loaded in recognition systems. The trained data are input which contains filtered letter from image. For example, Tamil letters such as are available in test dataset. The trained data are applied into deep convolution neural network process. The two dataset are created which contains test data with Tamil letter and second one for recognized input data or trained data. 15 thousands of letters are taken and 512 X 512 X 3 size deep convolution network is created with font and letters. As the result, 85% Tamil letters are recognized and 82% are tested using font. TensorFlow is used for testing the accuracy and success rate.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Hybrid deep learning model-based prediction of images related to cyberbullying
Autorzy:
Elmezain, Mahmoud
Malki, Amer
Gad, Ibrahim
Atlam, El-Sayed
Tematy:
cyberbullying
ResNet50
MobileNetV2
support vector machine
cyberprzemoc
maszyna wektorów wsparcia
Pokaż więcej
Wydawca:
Uniwersytet Zielonogórski. Oficyna Wydawnicza
Powiązania:
https://bibliotekanauki.pl/articles/2142490.pdf  Link otwiera się w nowym oknie
Opis:
Cyberbullying has become more widespread as a result of the common use of social media, particularly among teenagers and young people. A lack of studies on the types of advice and support available to victims of bullying has a negative impact on individuals and society. This work proposes a hybrid model based on transformer models in conjunction with a support vector machine (SVM) to classify our own data set images. First, seven different convolutional neural network architectures are employed to decide which is best in terms of results. Second, feature extraction is performed using four top models, namely, ResNet50, EfficientNetB0, MobileNet and Xception architectures. In addition, each architecture extracts the same number of features as the number of images in the data set, and these features are concatenated. Finally, the features are optimized and then provided as input to the SVM classifier. The accuracy rate of the proposed merged models with the SVM classifier achieved 96.05%. Furthermore, the classification precision of the proposed merged model is 99% in the bullying class and 93% in the non-bullying class. According to these results, bullying has a negative impact on students’ academic performance. The results help stakeholders to take necessary measures against bullies and increase the community’s awareness of this phenomenon.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Autism spectrum disorder detection in toddlers and adults using deep learning
Autorzy:
Abbas, Sidra
Ojo, Stephen
Krichen, Moez
Alamro, Meznah A.
Mihoub, Alaeddine
Vilcekova, Lucia
Tematy:
autism spectrum disorder
ASD
deep learning
feature fusion
feature prediction
healthcare technology
spektrum autyzmu
uczenie głębokie
fuzja funkcji
opieka zdrowotna
Pokaż więcej
Wydawca:
Uniwersytet Zielonogórski. Oficyna Wydawnicza
Powiązania:
https://bibliotekanauki.pl/articles/59123771.pdf  Link otwiera się w nowym oknie
Opis:
Autism spectrum disorder includes symptoms like anxiety, depressive disorders, and epilepsy because of its impact on relationships, learning, and employment. Since no confirmed treatment and diagnosis are available, the emphasis is on improving an individual’s capacities through symptom mitigation. This work investigates autism screening for adults and toddlers utilizing deep learning. We investigated models for feature prediction and fused these predictions with the original dataset to be trained with deep long short-term memory (DLSTM). Features are fused from the training and testing sets and then combined with the original dataset. Data analysis is carried out to detect anomalies and outliers, and a label encoding technique is utilized to convert the categorical data into numerical values. We hyper-tuned the DLSTM model parameters to optimize and assess significant outcomes. Experimental analysis and results revealed that the proposed approach worked better than the other techniques, yielding 99.9% accuracy for toddlers and 99% for adults.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Lung and colon cancer detection from CT images using deep learning
Autorzy:
Akinyemi, Joseph D.
Akinola, Akinkunle A.
Adekunle, Olajumoke O.
Adetiloye, Taiwo O.
Dansu, Emmanuel J.
Tematy:
cancer detection
efficientNet
CT images
healthcare
Pokaż więcej
Wydawca:
Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Instytut Informatyki Technicznej
Powiązania:
https://bibliotekanauki.pl/articles/3147620.pdf  Link otwiera się w nowym oknie
Opis:
Cancer is a deadly disease that has gained a reputation as a global health concern. Further, lung cancer has been widely reported as the most deadly cancer type globally, while colon cancer comes second. Meanwhile, early detection is one of the primary ways to prevent lung and colon cancer fatalities. To aid the early detection of lung and colon cancer, we propose a computer-aided diagnostic approach that employs a Deep Learning (DL) architecture to enhance the detection of these cancer types from Computed Tomography (CT) images of suspected body parts. Our experimental dataset (LC25000) contains 25 000 CT images of benign and malignant lung and colon cancer tissues. We used weights from a pre-trained DL architecture for computer vision, EfficientNet, to build and train a lung and colon cancer detection model. EfficientNet is a Convolutional Neural Network architecture that scales all input dimensions such as depth, width, and resolution at the same time. Our research findings showed detection accuracies of 99.63%, 99.50%, and 99.72% for training, validation, and test sets, respectively.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A Deep-Learning-Based Bug Priority Prediction Using RNN-LSTM Neural Networks
Autorzy:
Bani-Salameh, Hani
Sallam, Mohammed
Al shboul, Bashar
Tematy:
assigning
priority
bug tracking systems
bug priority
bug severity
closed-source
data mining
machine learning
ML
deep learning
RNN-LSTM
SVM
KNN
Pokaż więcej
Wydawca:
Politechnika Wrocławska. Oficyna Wydawnicza Politechniki Wrocławskiej
Powiązania:
https://bibliotekanauki.pl/articles/1818480.pdf  Link otwiera się w nowym oknie
Opis:
Context: Predicting the priority of bug reports is an important activity in software maintenance. Bug priority refers to the order in which a bug or defect should be resolved. A huge number of bug reports are submitted every day. Manual filtering of bug reports and assigning priority to each report is a heavy process, which requires time, resources, and expertise. In many cases mistakes happen when priority is assigned manually, which prevents the developers from finishing their tasks, fixing bugs, and improve the quality. Objective: Bugs are widespread and there is a noticeable increase in the number of bug reports that are submitted by the users and teams’ members with the presence of limited resources, which raises the fact that there is a need for a model that focuses on detecting the priority of bug reports, and allows developers to find the highest priority bug reports. This paper presents a model that focuses on predicting and assigning a priority level (high or low) for each bug report. Method: This model considers a set of factors (indicators) such as component name, summary, assignee, and reporter that possibly affect the priority level of a bug report. The factors are extracted as features from a dataset built using bug reports that are taken from closed-source projects stored in the JIRA bug tracking system, which are used then to train and test the framework. Also, this work presents a tool that helps developers to assign a priority level for the bug report automatically and based on the LSTM’s model prediction. Results: Our experiments consisted of applying a 5-layer deep learning RNN-LSTM neural network and comparing the results with Support Vector Machine (SVM) and K-nearest neighbors (KNN) to predict the priority of bug reports. The performance of the proposed RNN-LSTM model has been analyzed over the JIRA dataset with more than 2000 bug reports. The proposed model has been found 90% accurate in comparison with KNN (74%) and SVM (87%). On average, RNN-LSTM improves the F-measure by 3% compared to SVM and 15.2% compared to KNN. Conclusion: It concluded that LSTM predicts and assigns the priority of the bug more accurately and effectively than the other ML algorithms (KNN and SVM). LSTM significantly improves the average F-measure in comparison to the other classifiers. The study showed that LSTM reported the best performance results based on all performance measures (Accuracy = 0.908, AUC = 0.95, F-measure = 0.892).
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Big data significance in remote medical diagnostics based on deep learning techniques
Autorzy:
Kwaśniewska, A.
Giczewska, A.
Rumiński, J.
Tematy:
telemedicine
deep learning
multimedia databases
big data
telemedycyna
uczenie głębokie
multimedialne bazy danych
duże zbiory danych
Pokaż więcej
Wydawca:
Politechnika Gdańska
Powiązania:
https://bibliotekanauki.pl/articles/1940561.pdf  Link otwiera się w nowym oknie
Opis:
In this paper we discuss the evaluation of neural networks in accordance with medical image classification and analysis. We also summarize the existing databases with images which could be used for training deep models that can be later utilized in remote home-based health care systems. In particular, we propose methods for remote video-based estimation of patient vital signs and other health-related parameters. Additionally, potential challenges of using, storing and transferring sensitive patient data are discussed.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Towards a new deep learning algorithm based on GRU and CNN: NGRU
Autorzy:
Atassi, Abdelhamid
el Azami, Ikram
Tematy:
Convolutional Neural Network
CNN
Gated Recurrent Unit
GRU
SemEval
Twitter
word2vec
Keras
TensorFlow
Adadelta
Adam
soft-max
deep learning
Pokaż więcej
Wydawca:
Sieć Badawcza Łukasiewicz - Przemysłowy Instytut Automatyki i Pomiarów
Powiązania:
https://bibliotekanauki.pl/articles/2141895.pdf  Link otwiera się w nowym oknie
Opis:
This paper describes our new deep learning system based on a comparison between GRU and CNN. Initially we start with the first system which uses Convolutional Neural Network (CNN) which we will compare with the second system which uses Gated Recurrent Unit (GRU). And through this comparison we propose a new system based on the positive points of the two previous systems. Therefore, this new system will take the right choice of hyper-parameters recommended by the authors of both systems. At the final stage we propose a method to apply this new system to the dataset of different languages (used especially in socials networks).
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
An overview of deep learning techniques for short-term electricity load forecasting
Autorzy:
Adewuyi, Saheed
Aina, Segun
Uzunuigbe, Moses
Lawal, Aderonke
Oluwaranti, Adeniran
Tematy:
Short-term Load Forecasting
Deep Learning Architectures
RNN
LSTM
CNN
SAE
prognozowanie obciążenia krótkoterminowego
architektura głębokiego uczenia
Pokaż więcej
Wydawca:
Polskie Towarzystwo Promocji Wiedzy
Powiązania:
https://bibliotekanauki.pl/articles/117932.pdf  Link otwiera się w nowym oknie
Opis:
This paper presents an overview of some Deep Learning (DL) techniques applicable to forecasting electricity consumptions, especially in the short-term horizon. The paper introduced key parts of four DL architectures including the RNN, LSTM, CNN and SAE, which are recently adopted in implementing Short-term (electricity) Load Forecasting problems. It further presented a model approach for solving such problems. The eventual implication of the study is to present an insightful direction about concepts of the DL methods for forecasting electricity loads in the short-term period, especially to a potential researcher in quest of solving similar problems.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Anonymous traffic classification based on three-dimensional Markov image and deep learning
Autorzy:
Tang, Xin
Li, Huanzhou
Zhang, Jian
Tang, Zhangguo
Wang, Han
Cai, Cheng
Tematy:
anonymous network
traffic classification
three-dimensional Markov image
output self-attention
deep learning
sieć anonimowa
klasyfikacja ruchu
trójwymiarowy obraz Markowa
samouwaga wyjściowa
uczenie głębokie
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czasopisma i Monografie PAN
Powiązania:
https://bibliotekanauki.pl/articles/27311448.pdf  Link otwiera się w nowym oknie
Opis:
Illegal elements use the characteristics of an anonymous network hidden service mechanism to build a dark network and conduct various illegal activities, which brings a serious challenge to network security. The existing anonymous traffic classification methods suffer from cumbersome feature selection and difficult feature information extraction, resulting in low accuracy of classification. To solve this problem, a classification method based on three-dimensional Markov images and output self-attention convolutional neural network is proposed. This method first divides and cleans anonymous traffic data packets according to sessions, then converts the cleaned traffic data into three-dimensional Markov images according to the transition probability matrix of bytes, and finally inputs the images to the output self-attention convolution neural network to train the model and perform classification. The experimental results show that the classification accuracy and F1-score of the proposed method for Tor, I2P, Freenet, and ZeroNet can exceed 98.5%, and the average classification accuracy and F1-score for 8 kinds of user behaviors of each type of anonymous traffic can reach 93.7%. The proposed method significantly improves the classification effect of anonymous traffic compared with the existing methods.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Early detection of tuberculosis using hybrid feature descriptors and deep learning network
Autorzy:
Verma, Garima
Dixit, Sushil
Kumar, Ajay
Opis:
Purpose: To detect tuberculosis (TB) at an early stage by analyzing chest X-ray images using a deep neural network, and to evaluate the efficacy of proposed model by comparing it with existing studies. Material and methods: For the study, an open-source X-ray images were used. Dataset consisted of two types of images, i.e., standard and tuberculosis. Total number of images in the dataset was 4,200, among which, 3,500 were normal chest X-rays, and the remaining 700 X-ray images were of tuberculosis patients. The study proposed and simulated a deep learning prediction model for early TB diagnosis by combining deep features with hand-engineered features. Gabor filter and Canny edge detection method were applied to enhance the performance and reduce computation cost. Results: The proposed model simulated two scenarios: without filter and edge detection techniques and only a pre-trained model with automatic feature extraction, and filter and edge detection techniques. The results achieved from both the models were 95.7% and 97.9%, respectively. Conclusions: The proposed study can assist in the detection if a radiologist is not available. Also, the model was tested with real-time images to examine the efficacy, and was better than other available models.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Identifying selected diseases of leaves using deep learning and transfer learning models
Autorzy:
Mimi, Afsana
Zohura, Sayeda Fatema Tuj
Ibrahim, Muhammad
Haque, Riddho Ridwanul
Farrok, Omar
Jabid, Taskeed
Ali, Md Sawkat
Tematy:
convolutional neural network
transfer learning
leaf disease detection
image classification
Pokaż więcej
Wydawca:
Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Instytut Informatyki Technicznej
Powiązania:
https://bibliotekanauki.pl/articles/2204260.pdf  Link otwiera się w nowym oknie
Opis:
Leaf diseases may harm plants in different ways, often causing reduced productivity and, at times, lethal consequences. Detecting such diseases in a timely manner can help plant owners take effective remedial measures. Deficiencies of vital elements such as nitrogen, microbial infections and other similar disorders can often have visible effects, such as the yellowing of leaves in Catharanthus roseus (bright eyes) and scorched leaves in Fragaria ×ananassa (strawberry) plants. In this work, we explore approaches to use computer vision techniques to help plant owners identify such leaf disorders in their plants automatically and conveniently. This research designs three machine learning systems, namely a vanilla CNN model, a CNN-SVM hybrid model, and a MobileNetV2-based transfer learning model that detect yellowed and scorched leaves in Catharanthus roseus and strawberry plants, respectively, using images captured by mobile phones. In our experiments, the models yield a very promising accuracy on a dataset having around 4000 images. Of the three models, the transfer learning-based one demonstrates the highest accuracy (97.35% on test set) in our experiments. Furthermore, an Android application is developed that uses this model to allow end-users to conveniently monitor the condition of their plants in real time.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Sensor based cyber attack detections in critical infrastructures using deep learning algorithms
Autorzy:
Yilmaz, Murat
Catak, Ferhat Ozgur
Gul, Ensar
Tematy:
cyber security
engineering
critical infrastructures
industrial systems
information security
cyber attack detections
Pokaż więcej
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Powiązania:
https://bibliotekanauki.pl/articles/952946.pdf  Link otwiera się w nowym oknie
Opis:
The technology that has evolved with innovations in the digital world has also caused an increase in many security problems. Day by day the methods and forms of the cyberattacks began to become complicated, and therefore their detection became more difficult. In this work we have used the datasets which have been prepared in collaboration with Raymond Borges and Oak Ridge National Laboratories. These datasets include measurements of the Industrial Control Systems related to chewing attack behavior. These measurements include synchronized measurements and data records from Snort and relays with the simulated control panel. In this study, we developed two models using this datasets. The first is the model we call the DNN Model which was build using the latest Deep Learning algorithms. The second model was created by adding the AutoEncoder structure to the DNN Model. All of the variables used when developing our models were set parametrically. A number of variables such as activation method, number of hidden layers in the model, the number of nodes in the layers, number of iterations were analyzed to create the optimum model design. When we run our model with optimum settings, we obtained better results than related studies. The learning speed of the model has 100\% accuracy rate which is also entirely satisfactory. While the training period of the dataset containing about 4 thousand different operations lasts about 90 seconds, the developed model completes the learning process at the level of milliseconds to detect new attacks. This increases the applicability of the model in real world environment.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Classification and segmentation of periodontal cystfor digital dental diagnosis using deep learning
Autorzy:
Lakshmi, T. K.
Dheeba, J.
Tematy:
CNN
dental radiograph
deep learning
health care
machine transfer learning
periodontal cyst
predictive analytics
segmentation
U-Net
VGG16
rentgenowskie zdjęcie zębów
uczenie głębokie
opieka zdrowotna
uczenie się z transferu maszynowego
torbiel przyzębia
analityka predykcyjna
segmentacja
Pokaż więcej
Wydawca:
Instytut Podstawowych Problemów Techniki PAN
Powiązania:
https://bibliotekanauki.pl/articles/38700996.pdf  Link otwiera się w nowym oknie
Opis:
The digital revolution is changing every aspect of life by simulating the ways humansthink, learn and make decisions. Dentistry is one of the major fields where subsets ofartificial intelligence are extensively used for disease predictions. Periodontitis, the mostprevalent oral disease, is the main focus of this study. We propose methods for classifyingand segmenting periodontal cysts on dental radiographs using CNN, VGG16, and U-Net.Accuracy of 77.78% is obtained using CNN, and enhanced accuracy of 98.48% is obtainedthrough transfer learning with VGG16. The U-Net model also gives encouraging results.This study presents promising results, and in the future, the work can be extended withother pre-trained models and compared. Researchers working in this field can develop novelmethods and approaches to support dental practitioners and periodontists in decision-making and diagnosis and use artificial intelligence to bridge the gap between humansand machines.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Breast cancer nuclei segmentation and classification based on a deep learning approach
Autorzy:
Kowal, Marek
Skobel, Marcin
Gramacki, Artur
Korbicz, Józef
Tematy:
breast cancer
nuclei segmentation
image processing
nowotwór piersi
segmentacja jądra
przetwarzanie obrazu
Pokaż więcej
Wydawca:
Uniwersytet Zielonogórski. Oficyna Wydawnicza
Powiązania:
https://bibliotekanauki.pl/articles/1838197.pdf  Link otwiera się w nowym oknie
Opis:
One of the most popular methods in the diagnosis of breast cancer is fine-needle biopsy without aspiration. Cell nuclei are the most important elements of cancer diagnostics based on cytological images. Therefore, the first step of successful classification of cytological images is effective automatic segmentation of cell nuclei. The aims of our study include (a) development of segmentation methods of cell nuclei based on deep learning techniques, (b) extraction of some morphometric, colorimetric and textural features of individual segmented nuclei, (c) based on the extracted features, construction of effective classifiers for detecting malignant or benign cases. The segmentation methods used in this paper are based on (a) fully convolutional neural networks and (b) the marker-controlled watershed algorithm. For the classification task, seven various classification methods are used. Cell nuclei segmentation achieves 90% accuracy for benign and 86% for malignant nuclei according to the F-score. The maximum accuracy of the classification reached 80.2% to 92.4%, depending on the type (malignant or benign) of cell nuclei. The classification of tumors based on cytological images is an extremely challenging task. However, the obtained results are promising, and it is possible to state that automatic diagnostic methods are competitive to manual ones.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A deep-learning framework for enhancing habitat identification based on species composition
Autorzy:
Servajean, Maximilien
Leblanc, César
Šibík, Jozef
Stančić, Zvjezdana
Garbolino, Emmanuel
Uogintas, Domas
Lenoir, Jonathan
Moeslund, Jesper Erenskjold
Bergamini, Ariel
Jandt, Ute
Campos, Juan A.
Dengler, Jürgen
Golub, Valentin
Bonnet, Pierre
Čarni, Andraž
De Sanctis, Michele
Swacha, Grzegorz
Joly, Alexis
Pielech, Remigiusz
Argagnon, Olivier
Wohlgemuth, Thomas
Bonari, Gianmaria
Stanisci, Angela
Vassilev, Kiril
Biurrun, Idoia
Chytrý, Milan
Pérez-Haase, Aaron
Jansen, Florian
Lebedeva, Maria
Aćić, Svetlana
Ćušterevska, Renata
Opis:
Aims: The accurate classification of habitats is essential for effective biodiversity conservation. The goal of this study was to harness the potential of deep learning to advance habitat identification in Europe. We aimed to develop and evaluate models capable of assigning vegetation-plot records to the habitats of the European Nature Information System (EUNIS), a widely used reference framework for European habitat types. Location: The framework was designed for use in Europe and adjacent areas (e.g., Anatolia, Caucasus). Methods: We leveraged deep-learning techniques, such as transformers (i.e., models with attention components able to learn contextual relations between categorical and numerical features) that we trained using spatial k-fold cross-validation (CV) on vegetation plots sourced from the European Vegetation Archive (EVA), to show that they have great potential for classifying vegetation-plot records. We tested different network architectures, feature encodings, hyperparameter tuning and noise addition strategies to identify the optimal model. We used an independent test set from the National Plant Monitoring Scheme (NPMS) to evaluate its performance and compare its results against the traditional expert systems. Results: Exploration of the use of deep learning applied to species composition and plot-location criteria for habitat classification led to the development of a framework containing a wide range of models. Our selected algorithm, applied to European habitat types, significantly improved habitat classification accuracy, achieving a more than twofold improvement compared to the previous state-of-the-art (SOTA) method on an external data set, clearly outperforming expert systems. The framework is shared and maintained through a GitHub repository. Conclusions: Our results demonstrate the potential benefits of the adoption of deep learning for improving the accuracy of vegetation classification. They highlight the importance of incorporating advanced technologies into habitat monitoring. These algorithms have shown to be better suited for habitat type prediction than expert systems. They push the accuracy score on a database containing hundreds of thousands of standardized presence/absence European surveys to 88.74%, as assessed by expert judgment. Finally, our results showcase that species dominance is a strong marker of ecosystems and that the exact cover abundance of the flora is not required to train neural networks with predictive performances. The framework we developed can be used by researchers and practitioners to accurately classify habitats.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Deep learning-based automatic detection of tuberculosis disease in chest X-ray images
Autorzy:
Reiazi, Reza
Salehi, Mohammad
Sadighi, Nahid
Ghaffari, Hamed
Showkatian, Eman
Opis:
Purpose: To train a convolutional neural network (CNN) model from scratch to automatically detect tuberculosis (TB) from chest X-ray (CXR) images and compare its performance with transfer learning based technique of different pre-trained CNNs. Material and methods: We used two publicly available datasets of postero-anterior chest radiographs, which are from Montgomery County, Maryland, and Shenzhen, China. A CNN (ConvNet) from scratch was trained to automatically detect TB on chest radiographs. Also, a CNN-based transfer learning approach using five different pre-trained models, including Inception_v3, Xception, ResNet50, VGG19, and VGG16 was utilized for classifying TB and normal cases from CXR images. The performance of models for testing datasets was evaluated using five performances metrics, including accuracy, sensitivity/recall, precision, area under curve (AUC), and F1-score. Results: All proposed models provided an acceptable accuracy for two-class classification. Our proposed CNN architecture (i.e., ConvNet) achieved 88.0% precision, 87.0% sensitivity, 87.0% F1-score, 87.0% accuracy, and AUC of 87.0%, which was slightly less than the pre-trained models. Among all models, Exception, ResNet50, and VGG16 provided the highest classification performance of automated TB classification with precision, sensitivity, F1-score, and AUC of 91.0%, and 90.0% accuracy. Conclusions: Our study presents a transfer learning approach with deep CNNs to automatically classify TB and normal cases from the chest radiographs. The classification accuracy, precision, sensitivity, and F1-score for the detection of TB were found to be more than 87.0% for all models used in the study. Exception, ResNet50, and VGG16 models outperformed other deep CNN models for the datasets with image augmentation methods.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
On graph mining with deep learning: introducing model r for link weight prediction
Autorzy:
Hou, Yuchen
Holder, Lawrence B.
Tematy:
deep learning
neural networks
machine learning
graph mining
link weight prediction
predictive models
node embeddings
Pokaż więcej
Wydawca:
Społeczna Akademia Nauk w Łodzi. Polskie Towarzystwo Sieci Neuronowych
Powiązania:
https://bibliotekanauki.pl/articles/91884.pdf  Link otwiera się w nowym oknie
Opis:
Deep learning has been successful in various domains including image recognition, speech recognition and natural language processing. However, the research on its application in graph mining is still in an early stage. Here we present Model R, a neural network model created to provide a deep learning approach to the link weight prediction problem. This model uses a node embedding technique that extracts node embeddings (knowledge of nodes) from the known links’ weights (relations between nodes) and uses this knowledge to predict the unknown links’ weights. We demonstrate the power of Model R through experiments and compare it with the stochastic block model and its derivatives. Model R shows that deep learning can be successfully applied to link weight prediction and it outperforms stochastic block model and its derivatives by up to 73% in terms of prediction accuracy. We analyze the node embeddings to confirm that closeness in embedding space correlates with stronger relationships as measured by the link weight. We anticipate this new approach will provide effective solutions to more graph mining tasks
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A model of continual and deep learning for aspect based in sentiment analysis
Autorzy:
López, Dionis
Artigas-Fuentes, Fernando
Tematy:
continual learning
deep learning
catas
trophic forgetting
sentiment analysis
Pokaż więcej
Wydawca:
Sieć Badawcza Łukasiewicz - Przemysłowy Instytut Automatyki i Pomiarów
Powiązania:
https://bibliotekanauki.pl/articles/27314219.pdf  Link otwiera się w nowym oknie
Opis:
Sentiment analysis is a useful tool in several social and business contexts. Aspect sentiment classification is a subtask in sentiment analysis that gives information about features or aspects of people, entities, products, or services present in reviews. Different deep learning models that have been proposed to solve aspect sen‐ timent classification focus on a specific domain such as restaurant, hotel, or laptop reviews. However, there are few proposals for creating a single model with high performance in multiple domains. The continual learn‐ ing approach with neural networks has been used to solve aspect classification in multiple domains. However, avoiding low, aspect classification performance in contin‐ ual learning is challenging. As a consequence, potential neural network weight shifts in the learning process in different domains or datasets. In this paper, a novel aspect sentiment classification approach is proposed. Our approach combines a trans‐ former deep learning technique with a continual learning algorithm in different domains. The input layer used is the pretrained model Bidirectional Encoder Representations from Transformers. The experiments show the efficacy of our proposal with 78 % F1‐macro. Our results improve other approaches from the state‐of-the-art.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Estimating the distance to an object from grayscale stereo images using deep learning
Autorzy:
Kulawik, Joanna
Tematy:
estimating distance
stereo vision
convolutional neural network
deep learning
szacowanie odległości
widzenie stereoskopowe
konwolucyjne sieci neuronowe
uczenie głębokie
Pokaż więcej
Wydawca:
Politechnika Częstochowska. Wydawnictwo Politechniki Częstochowskiej
Powiązania:
https://bibliotekanauki.pl/articles/2202043.pdf  Link otwiera się w nowym oknie
Opis:
This article presents an innovative proposal for estimating the distance between an autonomous vehicle and an object in front of it. Such information can be used, for example, to support the process of controlling an autonomous vehicle. The primary source of information in research is monochrome stereo images. The images were made in compliance with the laws of the canonical order. The developed convolutional neural network model was used for the estimation. A proprietary dataset was developed for the experiments. The analysis was based on the phenomenon of disparity in stereo images. As a result of the research, a correctly trained model of the CNN network was obtained in six variants. High accuracy of distance estimation was achieved. This publication describes an original proposal for a hybrid blend of digital image analysis, stereo-vision, and deep learning for engineering applications.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Predictive modelling of turbofan engine components condition using machine and deep learning methods
Autorzy:
Matuszczak, Michał
Żbikowski, Mateusz
Teodorczyk, Andrzej
Tematy:
reliability
prognostics
deep learning
machine learning
gas turbine
turbofan engine
neural network
condition-based maintenance
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Polskie Naukowo-Techniczne Towarzystwo Eksploatacyjne PAN
Powiązania:
https://bibliotekanauki.pl/articles/1841686.pdf  Link otwiera się w nowym oknie
Opis:
The article proposes an approach based on deep and machine learning models to predict a component failure as an enhancement of condition based maintenance scheme of a turbofan engine and reviews currently used prognostics approaches in the aviation industry. Component degradation scale representing its life consumption is proposed and such collected condition data are combined with engines sensors and environmental data. With use of data manipulation techniques, a framework for models training is created and models' hyperparameters obtained through Bayesian optimization. Models predict the continuous variable representing condition based on the input. Best performed model is identified by detemining its score on the holdout set. Deep learning models achieved 0.71 MSE score (ensemble meta-model of neural networks) and outperformed significantly machine learning models with their best score at 1.75. The deep learning models shown their feasibility to predict the component condition within less than 1 unit of the error in the rank scale.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Remaining useful life prediction with insufficient degradation data based on deep learning approach
Autorzy:
Lyu, Yi
Jiang, Yijie
Zhang, Qichen
Chen, Ci
Tematy:
deep learning
remaining useful life
degradation data
data amplification
cycle-consistent generative adversarial network
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Polskie Naukowo-Techniczne Towarzystwo Eksploatacyjne PAN
Powiązania:
https://bibliotekanauki.pl/articles/2038109.pdf  Link otwiera się w nowym oknie
Opis:
Remaining useful life (RUL) prediction plays a crucial role in decision-making in conditionbased maintenance for preventing catastrophic field failure. For degradation-failed products, the data of performance deterioration process are the key for lifetime estimation. Deep learning has been proved to have excellent performance in RUL prediction given that the degradation data are sufficiently large. However, in some applications, the degradation data are insufficient, under which how to improve the prediction accuracy is yet a challenging problem. To tackle such a challenge, we propose a novel deep learning-based RUL prediction framework by amplifying the degradation dataset. Specifically, we leverage the cycle-consistent generative adversarial network to generate the synthetic data, based on which the original degradation dataset is amplified so that the data characteristics hidden in the sample space could be captured. Moreover, the sliding time window strategy and deep bidirectional long short-term memory network are employed to complete the RUL prediction framework. We show the effectiveness of the proposed method by running it on the turbine engine data set from the National Aeronautics and Space Administration. The comparative experiments show that our method outperforms a case without the use of the synthetically generated data.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A Study of Correlation between Fishing Activity and AIS Data by Deep Learning
Autorzy:
Shen, K. Y.
Chu, Y. J.
Chang, S. J.
Chang, S. M.
Tematy:
AIS Data
deep learning framework
learning methods
Recurrent Neural Network
(RNN)
Automatic Identification System
(AIS)
fishing operation
Pokaż więcej
Wydawca:
Uniwersytet Morski w Gdyni. Wydział Nawigacyjny
Powiązania:
https://bibliotekanauki.pl/articles/1841621.pdf  Link otwiera się w nowym oknie
Opis:
Previous researches on the prediction of fishing activities mainly rely on the speed over ground (SOG) as the referential attribute to determine whether the vessel is navigating or in fishing operation. Since more and more fishing vessels install Automatic Identification System (AIS) either voluntarily or under regulatory requirement, data collected from AIS in real time provide more attributes than SOG which may be utilized to improve the prediction. To be specific, the ships' trajectory patterns and the changes in course become available and should be considered. This paper aims to improve the accuracy in the identification of fishing activities. First, we do feature extraction from the AIS data of coastal waters around Taiwan and build a Recurrent Neural Network (RNN) model. Then, the activity data of fishing vessels are divided into fishing and non-fishing. Finally, based on the testing by feeding various fishing activity data, we can identify the fishing status automatically.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Zastosowanie uczenia głębokiego w tłumaczeniu komputerowym
Application of deep learning in computer translation
Autorzy:
Handzel, Zbigniew
Gajer, Mirosław
Grabiński, Tadeusz
Luty, Zbigniew
Tematy:
sztuczna inteligencja
przekład komputerowy
sieci neuronowe
uczenie głębokie
artificial intelligence
computer translation
neural
networks
deep learning
Pokaż więcej
Wydawca:
Wyższa Szkoła Ekonomii i Informatyki w Krakowie
Powiązania:
https://bibliotekanauki.pl/articles/2147416.pdf  Link otwiera się w nowym oknie
Opis:
Przekład komputerowy jest najstarszym i zarazem najbardziej doniosłym zagadnieniem zaliczanym do obszaru sztucznej inteligencji. Pomysł zastosowania komputerów do tłumaczenia tekstów zapisanych w języku naturalnym jest prawie tak stary, jak sam wynalazek komputera. Pierwotnie rzecz wydawała się łatwa do realizacji i oczekiwano, że za kilkanaście lat zawód tłumacza ostatecznie zaniknie, ponieważ tego rodzaju prace będą wykonywały wyłącznie maszyny cyfrowe. Potrzeba było jednak ponad 60 lat intensywnych badań, aby marzenie to mogło się urzeczywistnić w czasach nam współczesnych. Przełomem w badaniach nad przekładem komputerowym było zastosowanie technik obliczeniowych bazujących na sztucznych sieciach neuronowych z wykorzystaniem algorytmów uczenia głębokiego. W 2017 roku uruchomiony został serwis tłumaczeniowy DeepL, który jest programem komputerowym wykorzystującym uczenie głębokie w translacji automatycznej. Rozważany program zapewnia przekład o bardzo wysokiej jakości pomiędzy dowolnie wybraną parą spośród ponad 20 języków. Między innymi program ten umożliwia tłumaczenie z i na język polski. W artykule przedstawiono krótką historię badań nad przekładem komputerowym, omówiono główne trudności, które należało przezwyciężyć na drodze do budowy tłumaczy komputerowych, oraz omówiono podstawowe podejścia wykorzystywane w translacji automatycznej. Na zakończenie zaprezentowano interesujące wyniki eksperymentów przeprowadzonych z udziałem programu DeepL, które dowodzą jego bardzo wysokiej skuteczności w tłumaczeniu pomiędzy dowolnie wybraną parą języków, niezależnie od stopnia ich genetycznego pokrewieństwa.
Computer-aided translation is the oldest and at the same time the most prominent subject in the field of artificial intelligence. The idea of using computers to translate texts written in natural language is almost as old as the invention of the computer itself. At first it seemed easy to implement and it was expected that in a decade or so the profession of translator would finally disappear because only digital machines would do this kind of work. However, it took more than 60 years of intensive research for this dream to become a reality in modern times. A breakthrough in computer translation research was the application of computational techniques based on artificial neural networks using deep learning algorithms. In 2017, the translation service DeepL was launched, which is a computer program using deep learning in automatic translation. The program under consideration provides translation of very high quality between any pair of more than 20 languages. Among other things, the programme enables translation from and into Polish. The article presents a brief history of research on computer-aided translation, discusses the basic difficulties that had to be overcome on the way to building computer-aided translators, and discusses the basic approaches used in automatic translation. Finally, interesting results of experiments carried out with the program DeepL are presented, which prove its very high efficiency in translation between any pair of languages, regardless of the degree of their genetic affinity
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
A deep learning model for electricity demand forecasting based on a tropical data
Autorzy:
Adewuyi, Saheed A.
Aina, Segun
Oluwaranti, Adeniran I.
Tematy:
Electricity Demand Forecasting
STLF
Deep Learning Techniques
LSTM
CNN
MLP
prognozowanie zapotrzebowania na energię elektryczną
techniki głębokiego uczenia
Pokaż więcej
Wydawca:
Polskie Towarzystwo Promocji Wiedzy
Powiązania:
https://bibliotekanauki.pl/articles/118123.pdf  Link otwiera się w nowym oknie
Opis:
Electricity demand forecasting is a term used for prediction of users’ consumption on the grid ahead of actual demand. It is very important to all power stakeholders across levels. The power players employ electricity demand forecasting for sundry purposes. Moreover, the government’s policy on its market deregulation has greatly amplified its essence. Despite numerous studies on the subject using certain classical approaches, there exists an opportunity for exploration of more sophisticated methods such as the deep learning (DL) techniques. Successful researches about DL applications to computer vision, speech recognition, and acoustic computing problems are motivation. However, such researches are not sufficiently exploited for electricity demand forecasting using DL methods. In this paper, we considered specific DL techniques (LSTM, CNN, and MLP) to short-term load forecasting problems, using tropical institutional data obtained from a Transmission Company. We also test how accurate are predictions across the techniques. Our results relatively revealed models appropriateness for the problem.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Ship recognition and tracking system for intelligent ship based on deep learning framework
Autorzy:
Liu, B.
Wang, S. Z.
Xie, Z. X.
Zhao, J. S.
Li, M. F.
Tematy:
intelligent ship
deep learning framework
ship recognition system
ship tracking system
ship recognition and tracking system
intelligent navigation
autonomous ship
maritime safety
Pokaż więcej
Wydawca:
Uniwersytet Morski w Gdyni. Wydział Nawigacyjny
Powiązania:
https://bibliotekanauki.pl/articles/117419.pdf  Link otwiera się w nowym oknie
Opis:
Automatically recognizing and tracking dynamic targets on the sea is an important task for intelligent navigation, which is the prerequisite and foundation of the realization of autonomous ships. Nowadays, the radar is a typical perception system which is used to detect targets, but the radar echo cannot depict the target’s shape and appearance, which affects the decision-making ability of the ship collision avoidance. Therefore, visual perception system based on camera video is very useful for further supporting the autonomous ship navigational system. However, ship’s recognition and tracking has been a challenge task in the navigational application field due to the long distance detection and the ship itself motion. An effective and stable approach is required to resolve this problem. In this paper, a novel ship recognition and tracking system is proposed by using the deep learning framework. In this framework, the deep residual network and cross-layer jump connection policy are employed to extract the advanced ship features which help enhance the classification accuracy, thus improves the performance of the object recognition. Experimentally, the superiority of the proposed ship recognition and tracking system was confirmed by comparing it with state of-the-art algorithms on a large number of ship video datasets.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
VirulentHunter : deep learning-based virulence factor predictor illuminates pathogenicity in diverse microbial contexts
Autorzy:
Xiong, Xiangyi
Chmielarczyk, Agnieszka
Łabaj, Paweł
Shi, Tieliu
Zhang, Hao
Ouyang, Jian
Liu, Keyang
Chen, Chen
Xu, Yong
Wu, Jun
Różańska, Anna
Opis:
Virulence factors (VFs) are critical determinants of bacterial pathogenicity, but current homology-based identification methods often miss novel or divergent VFs, and many machine learning approaches neglect functional classification. Here, we present VirulentHunter, a novel deep learning framework that enable simultaneous VF identification and classification directly from protein sequences by leveraging the crucial step of fine-tuning pretrained protein language model. We curate a comprehensive VF database by integrating diverse public resources and expanding VF category annotations. Our benchmarking results demonstrate that VirulentHunter outperforms existing methods, particularly in identifying VFs lacking detectable homologs. Additionally, strain-level analysis using VirulentHunter highlights distinct pathogenicity profiles between Mycobacterium tuberculosis and Mycobacterium avium, revealing enrichment in VFs related to adherence, effector delivery systems, and immune modulation in M. tuberculosis, compared to biofilm formation and motility in M. avium. Furthermore, metagenomic profiling of gut microbiota from inflammatory bowel disease patient reveals a depletion of VFs associated with immune homeostasis. These results underscore the versatility of VirulentHunter as a powerful tool for VF analysis across diverse applications. To facilitate broader accessibility, we provide a freely accessible web service for VF prediction (http://www.unimd.org/VirulentHunter), accommodating protein sequences, genomes, and metagenomic data.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Zastosowanie uczenia głębokiego w tłumaczeniu komputerowym
Application of deep learning in computer translation
Autorzy:
Handzel, Zbigniew
Gajer, Mirosław
Wydawca:
Wydawnictwo Naukowe Wyższej Szkoły Ekonomii i Informatyki w Krakowie
Cytata wydawnicza:
Handzel, Z. Gajer, M. (2021). Zastosowanie uczenia głębokiego w tłumaczeniu komputerowym. Zeszyty Naukowe Wyższej Szkoły Ekonomii i Informatyki w Krakowie, 17, 72-93.
Opis:
Computer-aided translation is the oldest and at the same time the most prominent subject in the field of artificial intelligence. The idea of using computers to translate texts written in natural language is almost as old as the invention of the computer itself. At first it seemed easy to implement and it was expected that in a decade or so the profession of translator would finally disappear because only digital machines would do this kind of work. However, it took more than 60 years of intensive research for this dream to become a reality in modern times. A breakthrough in computer translation research was the application of computational techniques based on artificial neural networks using deep learning algorithms. In 2017, the translation service DeepL was launched, which is a computer program using deep learning in automatic translation. The program under consideration provides translation of very high quality between any pair of more than 20 languages. Among other things, the programme enables translation from and into Polish. The article presents a brief history of research on computer-aided translation, discusses the basic difficulties that had to be overcome on the way to building computer-aided translators, and discusses the basic approaches used in automatic translation. Finally, interesting results of experiments carried out with the program DeepL are presented, which prove its very high efficiency in translation between any pair of languages, regardless of the degree of their genetic affinity.
Dostawca treści:
Repozytorium Centrum Otwartej Nauki
Artykuł
Tytuł:
Classification of traffic over collaborative iot/cloud platforms using deep-learning recurrent LSTM
Autorzy:
Patil, Sonali A.
Raj, Arun L.
Tematy:
IoT
network traffic
machine learning
classification
cloud computing
Pokaż więcej
Wydawca:
Akademia Górniczo-Hutnicza im. Stanisława Staszica w Krakowie. Wydawnictwo AGH
Powiązania:
https://bibliotekanauki.pl/articles/2097958.pdf  Link otwiera się w nowym oknie
Opis:
The Internet of Things (IoT) and cloud-based collaborative platforms have emerged as new infrastructures over the recent decades. The classification of network traffic in terms of benign and malevolent traffic is indispensable for IoT/cloud-based collaborative platforms for optimally utilizing channel capac ity for transmitting benign traffic and blocking malicious traffic. The traffic classification mechanism should be dynamic and capable enough for classifying network traffic in a quick manner so that malevolent traffic can be identified at earlier stages and benign traffic can be speedily channelized to the destined nodes. In this paper, we present a deep-learning recurrent LSTM RNet-based technique for classifying traffic over IoT/cloud platforms using the Word2Vec approach. Machine-learning techniques (MLTs) have also been employed for comparing the performance of these techniques with the proposed LSTM RNet classification method. In the proposed research work, network traffic is clas sified into three classes: Tor-Normal, NonTor-Normal, and NonTor-Malicious traffic. The research outcome shows that the proposed LSTM RNet accurately classifies such traffic and also helps reduce network latency as well as enhance data transmission rates and network throughput.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Restoration of Remote Satellite Sensing Images using Machine and Deep Learning : a Survey
Autorzy:
Abdellaoui, Meriem
Benabdelkader, Souad
Assas, Ouarda
Tematy:
image restoration
remote sensing images
artificial intelligence
AI
machine learning
ML
deep learning
DL
convolutional neural network
CNN
Pokaż więcej
Wydawca:
Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Instytut Informatyki Technicznej
Powiązania:
https://bibliotekanauki.pl/articles/31339413.pdf  Link otwiera się w nowym oknie
Opis:
Remote sensing satellite images are affected by different types of degradation, which poses an obstacle for remote sensing researchers to ensure a continuous and trouble-free observation of our space. This degradation can reduce the quality of information and its effect on the reliability of remote sensing research. To overcome this phenomenon, the methods of detecting and eliminating this degradation are used, which are the subject of our study. The original aim of this paper is that it proposes a state of art of recent decade (2012-2022) on advances in remote sensing image restoration using machine and deep learning, identified by this survey, including the databases used, the different categories of degradation, as well as the corresponding methods. Machine learning and deep learning based strategies for remote sensing satellite image restoration are recommended to achieve satisfactory improvements.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Generating graphs using deep learning methods
Generowanie grafów z wykorzystaniem metod głębokiego uczenia
Autorzy:
Kasperczyk, Robert
Opis:
W niniejszej pracy przedstawiono klasyfikację oraz generowanie grafów przy wykorzystaniu geometrycznych grafowych sieci konwolucyjnych, które uwzględniają współrzędne wierzchołków. Opisywane podejście zostało zaproponowane przez grupę metod uczenia maszynowego GMUM.Realizacja tematu obejmuje autorską implementację wykorzystanej warstwy, stworzenie modelu klasyfikacji grafów, autoenkodera wariacyjnego (ang. VAE - Variational Autoencoder), a także zastosowanie podejścia adwersarialnego (ang. GAN - Generative Adversarial Network) i jego modyfikacji w postaci autoenkodera Wasserstein'a opartego o schemat uczenia GAN'a (ang. WAE-GAN - Wassernstein Autoencoder with GAN based penalty). Dodatkowo stworzono również środowisko do ewaluacji algorytmów uczenia maszynowego, dla danych w postaci struktur grafowych.Eksperymenty wykonano posługując się trzema zestawami danych, które obejmują reprezentacje grafów jako obrazy, grafy oraz związki chemiczne. Osiągnięte wyniki pokazują, że zaimplementowany model pozwala na uzyskanie porównywalnych z konkurencyjnymi modelami rezultatów w zadaniach klasyfikacyjnych oraz na generowanie poprawnych etykiet wierzchołków dla zbiorów danych MNIST oraz MNIST super piksele.
This work presents graph classification and generation using geometric graph convolutional networks which take nodes coordinates into account. For the first time, this approach was proposed by machine learning group of Jagiellonian University GMUM.Realisation of the topic includes implementation of the layer, graph classification model, creation of a Variational Autoencoder (VAE), Generative Adversarial Network (GAN) and it's modification - Wasserstein Autoencoder with GAN based penalty (WAE-GAN). Moreover, an environment for graph based machine learning algorithms evaluation has been created.Experiments were conducted using three datasets which include graph representation of images, graphs and chemical compounds. Achieved results show that the implemented models perform on a par with competitive solutions in classification tasks and generate correct nodes labels for MNIST and MNIST superpixels datasets.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Inne
Tytuł:
Additional Look into GAN-based Augmentation for Deep Learning COVID-19 Image Classification
Autorzy:
Fedoruk, Oleksandr
Klimaszewski, Konrad
Ogonowski, Aleksander
Kruk, Michał
Tematy:
computer vision
deep learning
image classification
generative adversarial network
medical imaging
Pokaż więcej
Wydawca:
Szkoła Główna Gospodarstwa Wiejskiego w Warszawie. Instytut Informatyki Technicznej
Powiązania:
https://bibliotekanauki.pl/articles/59122868.pdf  Link otwiera się w nowym oknie
Opis:
Data augmentation is a popular approach to overcome the insufficiency of training data for medical imaging. Classical augmentation is based on modification (rotations, shears, brightness changes, etc.) of the images from the original dataset. Another possible approach is the usage of Generative Adversarial Networks (GAN). This work is a continuation of the previous research where we trained StyleGAN2-ADA by Nvidia on the limited COVID-19 chest X-ray image dataset. In this paper, we study the dependence of the GAN-based augmentation performance on dataset size with a focus on small samples. Two datasets are considered, one with 1000 images per class (4000 images in total) and the second with 500 images per class (2000 images in total). We train StyleGAN2-ADA with both sets and then, after validating the quality of generated images, we use trained GANs as one of the augmentations approaches in multi-class classification problems. We compare the quality of the GAN-based augmentation approach to two different approaches (classical augmentation and no augmentation at all) by employing transfer learning-based classification of COVID-19 chest X-ray images. The results are quantified using different classification quality metrics and compared to the results from the previous article and literature. The GAN-based augmentation approach is found to be comparable with classical augmentation in the case of medium and large datasets but underperforms in the case of smaller datasets. The correlation between the size of the original dataset and the quality of classification is visible independently from the augmentation approach.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep Learning-based SNR Estimation for Multistage Spectrum Sensing in Cognitive Radio Networks
Autorzy:
Jeevangi, Sanjeevkumar
Jawaligi, Shivkumar
Patil, Vilaskumar
Tematy:
cognitive radio
improved NMF
LU-SLNO system
optimized CNN
spectrum sensing
Pokaż więcej
Wydawca:
Instytut Łączności - Państwowy Instytut Badawczy
Powiązania:
https://bibliotekanauki.pl/articles/2174451.pdf  Link otwiera się w nowym oknie
Opis:
Vacant frequency bands are used in cognitive radio (CR) by incorporating the spectrum sensing (SS) technique. Spectrum sharing plays a central role in ensuring the effectiveness of CR applications. Therefore, a new multi-stage detector for robust signal and spectrum sensing applications is introduced here. Initially, the sampled signal is subjected to SNR estimation by using a convolutional neural network (CNN). Next, the detection strategy is selected in accordance with the predicted SNR levels of the received signal. Energy detector (ED) and singular value-based detector (SVD) are the solutions utilized in the event of high SNR, whilst refined non-negative matrix factorization (MNMF) is employed in the case of low SNR. CNN weights are chosen via the Levy updated sea lion optimization (LU-SLNO) algorithm inspired by the traditional sea lion optimization (SLNO) approach. Finally, the outcomes of the selected detectors are added, offering a precise decision on spectrum tenancy and existence of the signal.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Constant Q-transform-based deep learning architecture for detection of obstructive sleep apnea
Autorzy:
Kandukuri, Usha Rani
Prakash, Allam Jaya
Patro, Kiran Kumar
Neelapu, Bala Chakravarthy
Tadeusiewicz, Ryszard
Pławiak, Paweł
Tematy:
sleep apnea
convolutional neural network
constant Q-transform
deep learning
single lead ECG signal
non apnea
obstructive sleep apnea
bezdech senny
sieć neuronowa konwolucyjna
uczenie głębokie
sygnał EKG
obturacyjny bezdech senny
Pokaż więcej
Wydawca:
Uniwersytet Zielonogórski. Oficyna Wydawnicza
Powiązania:
https://bibliotekanauki.pl/articles/24200694.pdf  Link otwiera się w nowym oknie
Opis:
Obstructive sleep apnea (OSA) is a long-term sleep disorder that causes temporary disruption in breathing while sleeping. Polysomnography (PSG) is the technique for monitoring different signals during the patient’s sleep cycle, including electroencephalogram (EEG), electromyography (EMG), electrocardiogram (ECG), and oxygen saturation (SpO2). Due to the high cost and inconvenience of polysomnography, the usefulness of ECG signals in detecting OSA is explored in this work, which proposes a two-dimensional convolutional neural network (2D-CNN) model for detecting OSA using ECG signals. A publicly available apnea ECG database from PhysioNet is used for experimentation. Further, a constant Q-transform (CQT) is applied for segmentation, filtering, and conversion of ECG beats into images. The proposed CNN model demonstrates an average accuracy, sensitivity and specificity of 91.34%, 90.68% and 90.70%, respectively. The findings obtained using the proposed approach are comparable to those of many other existing methods for automatic detection of OSA.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Deep learning-based CNC milling tool wear stage estimation with multi-signal analysis
Autorzy:
Karabacak, Yunus Emre
Tematy:
wear stage estimation
milling
convolutional neural network
time-frequency analysis
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Polskie Naukowo-Techniczne Towarzystwo Eksploatacyjne PAN
Powiązania:
https://bibliotekanauki.pl/articles/27312777.pdf  Link otwiera się w nowym oknie
Opis:
CNC milling machines are frequently used in the manufacturing of mechanical parts in the industry. One of the most important components of milling machines is the cutting tool. Monitoring the cutting tool wear is important for the reliability, continuity, and quality of production. Monitoring the tool and detecting the stage of wear are difficult processes. In this work, the convolutional neural network (CNN), which is a deep learning method in which the features are extracted by an inner process, was performed to detect the wear stages of the milling tool. These stages that define the total lifespan of the tool are known as initial wear (IW), steady-state wear (SSW), and accelerated wear (AW). Short Time Fourier Transform (STFT) was applied to signals, and signal spectrograms were used to train CNN models with different complex architectures. Vibration signals, acoustic emission signals, and motor current signals from The Nasa Ames Milling Dataset were used to obtain the spectrograms. Pre-trained CNNs (GoogleNet, AlexNet, ResNet-50, and EfficientNet-B0) detected the tool wear stage with varying accuracies. It has been seen that the time duration of model training increases as the size of the dataset grows and the network architecture becomes more complex. The recommended method has also been tested on the 2010 PHM Data Challenge Dataset. CNN shows promise for condition monitoring of milling operations and detecting tool wear stage.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Generating graphs using deep learning methods
Generowanie grafów z wykorzystaniem metod głębokiego uczenia
Autorzy:
Rzepecka, Katarzyna
Opis:
W niniejszej pracy przedstawiono klasyfikację oraz generowanie grafów przy wykorzystaniu geometrycznych grafowych sieci konwolucyjnych, które uwzględniają współrzędne wierzchołków. Opisywane podejście zostało zaproponowane przez grupę metod uczenia maszynowego GMUM.Realizacja tematu obejmuje autorską implementację wykorzystanej warstwy, stworzenie modelu klasyfikacji grafów, autoenkodera wariacyjnego (ang. VAE - Variational Autoencoder), a także zastosowanie podejścia adwersarialnego (ang. GAN - Generative Adversarial Network) i jego modyfikacji w postaci autoenkodera Wasserstein'a opartego o schemat uczenia GAN'a (ang. WAE-GAN - Wassernstein Autoencoder with GAN based penalty). Dodatkowo stworzono również środowisko do ewaluacji algorytmów uczenia maszynowego, dla danych w postaci struktur grafowych.Eksperymenty wykonano posługując się trzema zestawami danych, które obejmują reprezentacje grafów jako obrazy, grafy oraz związki chemiczne. Osiągnięte wyniki pokazują, że zaimplementowany model pozwala na uzyskanie porównywalnych z konkurencyjnymi modelami rezultatów w zadaniach klasyfikacyjnych oraz na generowanie poprawnych etykiet wierzchołków dla zbiorów danych MNIST oraz MNIST super piksele.
This work presents graph classification and generation using geometric graph convolutional networks which take nodes coordinates into account. For the first time, this approach was proposed by machine learning group of Jagiellonian University GMUM.Realisation of the topic includes implementation of the layer, graph classification model, creation of a Variational Autoencoder (VAE), Generative Adversarial Network (GAN) and it's modification - Wasserstein Autoencoder with GAN based penalty (WAE-GAN). Moreover, an environment for graph based machine learning algorithms evaluation has been created.Experiments were conducted using three datasets which include graph representation of images, graphs and chemical compounds. Achieved results show that the implemented models perform on a par with competitive solutions in classification tasks and generate correct nodes labels for MNIST and MNIST superpixels datasets.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Inne
Tytuł:
Deep learning (pogłębianie procesu uczenia się) z perspektywy analizy potrzeb studentów języka angielskiego jako obcego
Deep Learning from the Perspective of Needs Analysis of Students of English as a Foreign Language
Autorzy:
Papaja, Katarzyna
Świątek, Artur
Mielnik, Kamil
Tematy:
Deep Learning
process
teacher
student
foreign language
personalisation
transformation
deep learning
pogłębione uczenie się
proces
nauczyciel
uczeń
język obcy
personalizacja
transformacja
Pokaż więcej
Wydawca:
Ateneum - Akademia Nauk Stosowanych w Gdańsku
Powiązania:
https://bibliotekanauki.pl/articles/1398073.pdf  Link otwiera się w nowym oknie
Opis:
Although the term Deep Learning does not seem to be a new term in language learning, it attracted relatively little attention until just a few years ago. Different fields of study show that Deep Learning leverages a sophisticated process to learn multiple levels of abstraction from the data; however, in languages, the term has been widely accepted as the key concept in the transformation and personalisation of the learning process. In this paper, we take the definition of Deep Learning, and we corroborate the theories by use of the study which aims to assess the needs of students in the context of language exercises, resources as well as tools and modern technological solutions. A proper understanding of Deep Learning is necessary to examine the potential benefits for students and the broadly defined society. Therefore, the essence of the research is to obtain the answers to what is important in the education of modern foreign languages and also what the teacher’s role is. A quantitative study was conducted on 441 students of English Philology. The results of the needs analysis of foreign language students allow for a greater understanding of their expectations towards themselves and their teachers; additionally, to answer the question about what kind of education recipients they are and whether they are active participants in the whole educational process.
Choć termin pogłębionego procesu uczenia się (deep learning) nie wydaje się być terminem nowym w nauczaniu języków, do niedawna przyciągnął stosunkowo niewiele uwagi naukowców. W wielu językach jednak termin ten został powszechnie zaakceptowany jako kluczowa koncepcja transformacji i personalizacji procesu uczenia się. W niniejszym artykule prezentujemy definicję deep learning i potwierdzamy teorię poprzez badanie, którego celem jest ocena potrzeb uczniów w kontekście ćwiczeń językowych, zasobów, a także narzędzi i nowoczesnych rozwiązań technologicznych. Prawidłowe zrozumienie pogłębionego uczenia się jest konieczne, aby zbadać potencjalne korzyści wynikające z niego dla studentów i szeroko rozumianego społeczeństwa. Dlatego też istotą prowadzonych badań jest uzyskanie odpowiedzi na pytanie, co jest ważne w dydaktyce współczesnych języków obcych, a także jaka jest rola nauczyciela w tym zakresie. Wyniki analiz potrzeb uczniów języków obcych pozwalają uzyskać wiedzę na temat ich oczekiwań wobec siebie samych oraz wobec nauczycieli, a także odpowiedzieć na pytanie, jakiego rodzaju odbiorcami edukacji są młodzi uczący się i czy aktywnie partycypują w globalnym procesie kształcenia.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Early detection of major diseases in turmeric plant using improved deep learning algorithm
Autorzy:
Devisurya, V.
Devi Priya, R.
Anitha, N.
Tematy:
artificial intelligence
computer vision
turmeric leaf diseases detection
sztuczna inteligencja
wizja komputerowa
wykrywanie chorób liści kurkumy
Pokaż więcej
Wydawca:
Polska Akademia Nauk. Czytelnia Czasopism PAN
Powiązania:
https://bibliotekanauki.pl/articles/2173642.pdf  Link otwiera się w nowym oknie
Opis:
Turmeric is affected by various diseases during its growth process. Not finding its diseases at early stages may lead to a loss in production and even crop failure. The most important thing is to accurately identify diseases of the turmeric plant. Instead of using multiple steps such as image pre-processing, feature extraction, and feature classification in the conventional method, the single-phase detection model is adopted to simplify recognizing turmeric plant leaf diseases. To enhance the detection accuracy of turmeric diseases, a deep learning-based technique called the Improved YOLOV3-Tiny model is proposed. To improve detection accuracy than YOLOV3-tiny, this method uses residual network structure based on the convolutional neural network in particular layers. The results show that the detection accuracy is improved in the proposed model compared to the YOLOV3-Tiny model. It enables anyone to perform fast and accurate turmeric leaf diseases detection. In this paper, major turmeric diseases like leaf spot, leaf blotch, and rhizome rot are identified using the Improved YOLOV3-Tiny algorithm. Training and testing images are captured during both day and night and compared with various YOLO methods and Faster R-CNN with the VGG16 model. Moreover, the experimental results show that the Cycle-GAN augmentation process on turmeric leaf dataset supports much for improving detection accuracy for smaller datasets and the proposed model has an advantage of high detection accuracy and fast recognition speed compared with existing traditional models.
Dostawca treści:
Biblioteka Nauki
Artykuł
Tytuł:
Differentiation of carcinosarcoma from endometrial carcinoma on magnetic resonance imaging using deep learning
Autorzy:
Satoh, Toyomi
Saida, Tsukasa
Ishiguro, Toshitaka
Sakai, Masafumi
Hoshiai, Sodai
Mori, Kensaku
Nakajima, Takahito
Urushibara, Aiko
Opis:
Purpose: To verify whether deep learning can be used to differentiate between carcinosarcomas (CSs) and endometrial carcinomas (ECs) using several magnetic resonance imaging (MRI) sequences. Material and methods: This retrospective study included 52 patients with CS and 279 patients with EC. A deep-learning model that uses convolutional neural networks (CNN) was trained with 572 T2-weighted images (T2WI) from 42 patients, 488 apparent diffusion coefficient of water maps from 33 patients, and 539 fat-saturated contrast-enhanced T1-weighted images from 40 patients with CS, as well as 1612 images from 223 patients with EC for each sequence. These were tested with 9-10 images of 9-10 patients with CS and 56 images of 56 patients with EC for each sequence, respectively. Three experienced radiologists independently interpreted these test images. The sensitivity, specificity, accuracy, and area under the receiver operating characteristic curve (AUC) for each sequence were compared between the CNN models and the radiologists. Results: The CNN model of each sequence had sensitivity 0.89-0.93, specificity 0.44-0.70, accuracy 0.83-0.89, and AUC 0.80-0.94. It also showed an equivalent or better diagnostic performance than the 3 readers (sensitivity 0.43-0.91, specificity 0.30-0.78, accuracy 0.45-0.88, and AUC 0.49-0.92). The CNN model displayed the highest diagnostic performance on T2WI (sensitivity 0.93, specificity 0.70, accuracy 0.89, and AUC 0.94). Conclusions: Deep learning provided diagnostic performance comparable to or better than experienced radiologists when distinguishing between CS and EC on MRI.
Dostawca treści:
Repozytorium Uniwersytetu Jagiellońskiego
Artykuł
Tytuł:
Career track prediction using deep learning model based on discrete series of quantitative classification
Autorzy:
Hernandez, Rowell
Atienza, Robert
Tematy:
track prediction
deep learning
education
przewidywanie torów
głębokie uczenie
edukacja
Pokaż więcej
Wydawca:
Polskie Towarzystwo Promocji Wiedzy
Powiązania:
https://bibliotekanauki.pl/articles/1956033.pdf  Link otwiera się w nowym oknie
Opis:
In this paper, a career track recommender system was proposed using Deep Neural Network model. This study aims to assist guidance counselors in guiding their students in the selection of a suitable career track. It is because a lot of Junior High school students experienced track uncertainty and there are instances of shifting to another program after learning they are not suited for the chosen track or course in college. In dealing with the selection of the best student attributes that will help in the creation of the predictive model, the feature engineering technique is used to remove the irrelevant features that can affect the performance of the DNN model. The study covers 1500 students from the first to the third batch of the K-12 curriculum, and their grades from 11 subjects, sex, age, number of siblings, parent’s income, and academic strand were used as attributes to predict their academic strand in Senior High School. The efficiency and accuracy of the algorithm depend upon the correctness and quality of the collected student’s data. The result of the study shows that the DNN algorithm performs reasonably well in predicting the academic strand of students with a predic-tion accuracy of 83.11%. Also, the work of guidance counselors became more efficient in handling students’ concerns just by using the proposed system. It is concluded that the recommender system serves as a decision tool for counselors in guiding their stu-dents to determine which Senior High School track is suitable for students with the utilization of the DNN model.
Dostawca treści:
Biblioteka Nauki
Artykuł

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