168 results
Search Results
2. High-end image classification on the dogs vs. cats dataset using convolutional neural network.
- Author
-
Chutani, P., Gulati, S., and Arora, N.
- Subjects
CONVOLUTIONAL neural networks ,ARTIFICIAL neural networks ,DATA augmentation ,MACHINE learning ,DEEP learning - Abstract
The enormous benefits and applications of Image classification and recognition are umpteen. Machine learning algorithms and Deep Neural Networks are like windfall to fathom the objective proficiently in streamlined manner. The prevalent improvement in this technology is that these networks do not call for any prior blueprints in terms of algorithms as prerequisites. The presented paper is an attempt to create a Convolutional Neural Network from scratch to classify the images from the well-known dataset – Cats and Dogs into their relevant baskets. Manifold open source accessible approaches to amplify the efficiency of the network are no more onerous. Further, data augmentation technique boosts the efficiency tremendously by extending the dataset with reoriented features from the same images. To untangle the same problem, Transfer Learning is also a compelling technique in which all the layers, neurons in each layer, weights of each neuron and all other parameters are predefined and we can amend the output layer as per the classes in the respective problem statement. In the present paper, we have tried to obtain a comparable efficiency with a significant reduction in parameters. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
3. Brain-inspired learning in artificial neural networks: A review.
- Author
-
Schmidgall, Samuel, Ziaei, Rojin, Achterberg, Jascha, Kirsch, Louis, Hajiseyedrazi, S. Pardis, and Eshraghian, Jason
- Subjects
ARTIFICIAL neural networks ,MACHINE learning ,IMAGE processing ,COMPUTATIONAL biology ,DATA integration - Abstract
Artificial neural networks (ANNs) have emerged as an essential tool in machine learning, achieving remarkable success across diverse domains, including image and speech generation, game playing, and robotics. However, there exist fundamental differences between ANNs' operating mechanisms and those of the biological brain, particularly concerning learning processes. This paper presents a comprehensive review of current brain-inspired learning representations in artificial neural networks. We investigate the integration of more biologically plausible mechanisms, such as synaptic plasticity, to improve these networks' capabilities. Moreover, we delve into the potential advantages and challenges accompanying this approach. In this review, we pinpoint promising avenues for future research in this rapidly advancing field, which could bring us closer to understanding the essence of intelligence. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF
4. Review on disease detection of plants using image processing and machine learning techniques.
- Author
-
Santhosh Kumar, P., Balakrishna, R., and Vinod Kumar, K.
- Subjects
IMAGE processing ,ARTIFICIAL neural networks ,CROPS ,ARTIFICIAL intelligence ,BACK propagation ,DEEP learning ,MACHINE learning ,PLANT diseases - Abstract
Agriculture is essential for everyone to promote sustainable development, the farming, which combines image processing, artificial intelligence, Deep learning, and Internet of Things (IOT). World population incensing every day. Due to the rising demand in the Agriculture industry, the need to collectively improve a plants and growth its field is very useful. In this paper, it is important to maintain the crop during its initial time, and also at period of harvesting. The image processing and artificial networks are used as a different techniques to maintain the detecting the diseases on the leaves and correct time to harvesting. When we take images with help of drones, the images are divided and changed to disease described three things vectors namely the first one is color, one more is texture and morphology. The vectors morphology gives 95% accuracy and its give more compare to other two vector features. This research paper proposed effective and useful algorithms for detection of disease with help implementation of Artificial Neural network algorithms using MATLAB. Detection of leaf or plant diseases with some manual techniques are requires a lot of work by maintaining a huge farm of crops, and it's very early stages it detects different types of symptoms to different diseases on plants, when the displayed on crop leaves. In this research paper survey on various disease classification techniques that can be for plant leaves diseases detection. For this purpose Artificial Intelligence, Neural network algorithms and back propagation techniques for adjustment of training data sets. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
5. Artificial intelligence based learning for wireless application – A survey.
- Author
-
Raja, L., Velmurugan, S., Shanthi, G., and Nirmala, S.
- Subjects
ARTIFICIAL intelligence ,ARTIFICIAL neural networks ,DEEP learning ,NEXT generation networks ,MACHINE learning ,COMPUTER network traffic - Abstract
The wireless networks of the future generation are evolved as complex systems due to broadening in service prerequisites, application heterogeneity and networking of gadgets. In recent times, the major step forward of machine learning techniques is deep learning. However, in wireless/heterogeneous networks, the deep learning application for network traffic control is relatively new. The key disputes in the wireless backbone networks since the advancement of wireless networks are resource allocation and efficient network traffic control such as routing. In larger scale of networks and tangled radio environments, the method of adding intelligence to wireless networks is achieved by Deep Learning (DL). The tangled wireless networks supported with several nodes and their quality of variable link can be investigated by Deep Learning. This paper is presented with a perception of harnessing the next generation communication networks with the artificial neural networks. This work helps the readers to explore the unsolved issues to pursue their research and deeply understand the wireless network design with DL based state of the art facilities. In this work, we integrate the deep learning and wireless networking research with a widespread survey. Finally, this paper summarizes the disputes and benefits of acquiring ML and AI for next generation wireless systems. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
6. Vision transformer based Devanagari character recognition.
- Author
-
Kumar, Shailendra, Chopra, Abhinav, Jain, Sambhav, and Arora, Sarthak
- Subjects
TRANSFORMER models ,ARTIFICIAL neural networks ,PATTERN recognition systems ,HANDWRITING recognition (Computer science) ,COMPUTER vision ,MACHINE learning - Abstract
Devanagari is an ancient script that is used to write Hindi, Nepali, Marathi, Maithili, Awadhi, Newari, and Bhojpuri, among other Indo-Aryan languages. Thousands of individuals in India use this script to write documents in Marathi and Hindi. Indian mythology is based on this script. Because of the script's prominence, handwritten Devanagari character identification has grown in popularity over time. Handwritten recognition of languages such as English has received a lot of attention, but Indian languages written in the Devanagari script are also a rich source of information. Most of the work on this problem statement has been done either using deep neural networks like CNN at its heart coupled with other machine learning techniques like SVM,Random Forest etc. In this paper we are utilising a recently introduced transformer model for computer vision known as Vision Transformer for the task of Devanagari Character Recognition. We have also compared our model with various pretrained CNN-based architectures like ResNet50,VGG16 and InceptionV3 and ViT has outperformed these models both on DHCD dataset and the modified slightly more complex version of it with accuracy scores of 99.68% on the original testing dataset of the DHCD dataset and accuracy score of 96.55% on the modified(blurred) slightly more complex version of the original testing dataset. The ViT model thus generalized better than standard CNN-based models on the problem of Devanagari Character recognition. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
7. Enhanced all-optical vector atomic magnetometer enabled by artificial neural network.
- Author
-
Qin, Jianan, Xu, Jinxin, Jiang, Zhiyuan, and Qu, Jifeng
- Subjects
- *
ARTIFICIAL neural networks , *MACHINE learning , *MAGNETOMETERS , *SIGNALS & signaling , *ANGLES - Abstract
This paper reports an all-optical vector magnetometer enhanced by a machine learning model. Using a dual probing beam setup, spin projections in both probe directions are simultaneously detected. Vector information is directly obtained from the measured phases of spin projection signals. To enhance the measurement accuracy and mitigate the dead zone effect, we introduce an artificial neural network (ANN) to link the phase signals to the field direction. With the addition of amplitude information to the ANN input, the average angle error is reduced to less than 0.3 ° within a hemisphere. Furthermore, this configuration demonstrates a field angle sensitivity of better than 30 μ rad / Hz 1 / 2 . [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF
8. Rapid detection and discrimination of plant leaves using laser-induced breakdown spectroscopy.
- Author
-
Cai, Jinzhu, Wu, Tianzhuang, Chen, Yu, Yang, Siyuan, Zhang, Zhirong, and Liu, Yuzhu
- Subjects
ARTIFICIAL neural networks ,CONVOLUTIONAL neural networks ,LASER-induced breakdown spectroscopy ,BACK propagation ,MACHINE learning - Abstract
The wide diversity of species and the remarkable variation in morphological features that allow plants to adapt to a wide range of terrestrial environments is a fact that highlights the fundamental and crucial role of plants in the field of biodiversity studies. Currently, research on leaf classification is limited and in its early stages. A novel classification system based on laser-induced breakdown spectroscopy (LIBS) technology was proposed in this paper, integrated with machine learning for real-time, in situ detection and analysis of leaves. Four representative leaf samples—Ilex chinensis, Camellia japonica, Cinnamomum camphora, and Osmanthus fragrans—were subjected to spectral analysis and machine learning techniques. Spectral analysis revealed distinct spectral lines corresponding to elements such as Ca, Al, Mg, Na, and Fe, alongside common elements including C, N, and O. Principal component analysis (PCA) was employed to reduce the dimensionality of the spectral data, and the first 13 principal components used in this study captured 98.76% of the total variance. Following this, support vector machine (SVM), backpropagation artificial neural network and convolutional neural network (CNN) algorithms were applied for machine learning on the principal components to develop leaf recognition classification models. Through comparison, the CNN algorithm, boasting a classification accuracy of up to 94.44%, was ultimately selected. The models established by SVM and back propagation artificial neural network achieved accuracy of only 83.33% and 90.00%, respectively. The results suggest that integrating LIBS with machine learning is an effective and precise approach for leaf classification, offering promising applications in biodiversity research. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF
9. Conquering fashion MNIST with CNNs using computer vision by pretrained models: VGG19 and RESNET50.
- Author
-
Venkataravanappa, Viswanatha, Chowdappa, Ramachandra Ankathattahalli, Shamanna, Madhukara, Krishnappa, Manjula, Mariyappa, Bavitesh, and Singh, Abhishek Kumar
- Subjects
- *
CONVOLUTIONAL neural networks , *ARTIFICIAL neural networks , *MACHINE learning , *COMPUTER vision , *FASHION - Abstract
This paper delves into the training and testing of two pre-trained models of a Convolutional Neural Network (CNN) to classify images of clothing from the Fashion MNIST dataset and determine the classification accuracy and performance of both the models. The Fashion MNIST dataset is a collection of 60,000 28x28 grayscale images of 10 different types of clothing. The images are well-labeled and comparatively easy to classify, making them a good starting point for learning about CNNs. The article starts with a summary of the Fashion MNIST dataset. There are 60,000 training photos and ten thousand test ones in the dataset. T-shirt/top, Trouser, Pullover, Dress, Coat, Sandal, Shirt, Sneaker, Bag and Ankle boot are the pictures' ten constituent classes. The CNN architecture used in the tests is described in the next part of the study. CNN architecture is made up of two convolutional layers, two max-pooling layers, and two fully linked layers. The convolutional layers employ 32 3x3 filters, whereas the max pooling layers employ a pool size of 2x2. There are 128 neurons in the fully linked layer and 10 neurons in the unconnected layer. The Adam optimizer is chosen to train the CNN framework utilizing the learning rate of 1e-5. The machine learning algorithm underwent training for 30 epochs with a batch size of 64. The experiments' shortcomings are highlighted in the conclusion section. It is certainly possible that the CNN model would not perform as well on other datasets. Another limitation is that the experiments were conducted only for a set of hyperparameters. The outcomes of the experiments are presented in the results section. The training accuracy of the Convolutional Neural Network models VGG19 and Resnet50 are 94.58 and 99.41 respectively. The testing accuracies for VGG19 and Resnet50 models are 91.92 and 90.25 respectively. The inference latencies for both cases were found to be 5.439 and 3.522 respectively. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF
10. Evaluation of performance over various pre-trained deepconvolutional neural network models for co-saliency detection problem.
- Author
-
Mangal, Anuj, Garg, Hitendra, and Bhatnagar, Charul
- Subjects
ARTIFICIAL neural networks ,MACHINE learning ,DEEP learning ,COMPUTER vision ,VISUAL fields - Abstract
It's possible that our Human Visual System (HVS) only sees a particular section of an image instead of the whole picture. This phenomenon is a trending and demanding topic in the fields of computer vision research field. For predicting co-salient features many deep learning algorithms have recently been used. This paper examines the visual saliency prediction ability of five cutting-edge deep CNN models namedResnet-50, InceptionResNet-v2, VGG-16, Xception and MobileNet-v2. By using the SALICON dataset, we used five deep learning pre trained models that are used to suggestco-saliency maps on four popular datasets: DUT-OMRON, TORONTO, CoSOD3k, MIT103. According to the data, the ResNet-50 model outperforms in obtaining desirable results and assisted in giving more accurate results that closely resembles the desired output. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
11. Fuzzy Neural Network-based Fetal Health Monitoring using Cardiotocography Data.
- Author
-
Chang-Wook Han
- Subjects
FETAL monitoring ,FETAL heart rate monitoring ,FUZZY neural networks ,GENETIC algorithms ,MACHINE learning ,ARTIFICIAL neural networks - Abstract
Fuzzy neural networks have been widely applied in the medical field. In this paper, we apply cascade architectures of fuzzy neural networks to monitoring fetal health using Cardiotocography data. Cascade architectures of neural networks can select reduced size of input subspace by selecting useful inputs. For the optimization of the input subspace and the structure of the cascade architectures of fuzzy neural networks, genetic algorithms and gradient decent method are used. To verify the applicability of the proposed method, Cardiotocography data available on the Machine Learning Repository site at the University of California at Irvine is used. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
12. Employing artificial bee colony algorithm to optimize the artificial neural network in heart disease prediction.
- Author
-
Asaad, Manal Mohammed Othman Farea, Wahid, Juliana, and Rahmat, Abdul Razak
- Subjects
- *
BEES algorithm , *HEART diseases , *HONEYBEES , *ARTIFICIAL neural networks , *MACHINE learning , *MACHINE tools , *HONEY - Abstract
Heart disease forecasting is a key issue in the clinical data analysis field. Artificial Neural Network (ANN) is a machine learning tool that can help doctors diagnose heart diseases more accurately and quickly. However, the design of ANN is complicated because it needs to identify optimal weight values and suitable network structures. This paper aims to optimize the weights of ANN for forecasting the existence of heart disease among humans by using Artificial Bee Colony (ABC) algorithm to train ANN and select optimal network weights. The experimentations are carried out on the UC Irvine Machine Learning Repository (UCI) heart disease datasets and tested by the MATLAB software machine learning tools. The accuracy, specificity, sensitivity (recall), precision, and F-score of the ANN trained by the ABC model (ABC-ANN) and ANN trained by ABC with backpropagation (ABC-BpNN) model are investigated. Between ABC-ANN and ABC-BpNN, it was shown that the ABC-BpNN model has a better predicting ability and could reach considerably higher accuracy with 90% and 90.9%, 88.9%, 90.8% and 90.9% for precision, specificity, recall, and F-score, respectively. The findings showed significant enhancements compared with previous studies that have utilized the same dataset. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF
13. On the incorporation of conservation laws in machine learning tabulation of kinetics for reacting flow simulation.
- Author
-
Readshaw, Thomas, Jones, W. P., and Rigopoulos, Stelios
- Subjects
FLOW simulations ,CONSERVATION laws (Physics) ,MACHINE learning ,ARTIFICIAL neural networks ,CHEMICAL kinetics - Abstract
Tabulation of chemical mechanisms with artificial neural networks (ANNs) offers significant speed benefits when computing the real-time integration of reaction source terms in turbulent reacting flow simulations. In such approaches, the ANNs should be physically consistent with the reaction mechanism by conserving mass and chemical elements, as well as obey the bounds of species mass fractions. In the present paper, a method is developed for satisfying these constraints to machine precision. The method can be readily applied to any reacting system and appended to the existing ANN architectures. To satisfy the conservation laws, certain species in a reaction mechanism are selected as residual species and recalculated after ANN predictions of all of the species have been made. Predicted species mass fractions are set to be bounded. While the residual species mass fractions are not guaranteed to be non-negative, it is shown that negative predictions can be avoided in almost all cases and easily rectified if necessary. The ANN method with conservation is applied to one-dimensional laminar premixed flame simulations, and comparisons are made with simulations performed with direct integration (DI) of chemical kinetics. The ANNs with conservation are shown to satisfy the conservation laws for every reacting point to machine precision and, furthermore, to provide results in better agreement with DI than ANNs without conservation. It is, thus, shown that the proposed method reduces accumulation of errors and positively impacts the overall accuracy of the ANN prediction at negligible additional computational cost. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
14. Predicting the effect of inertia, rotation, and magnetic field on the onset of convection in a bidispersive porous medium using machine learning techniques.
- Author
-
Singh, Mahesh, Ragoju, Ravi, Shiva Kumar Reddy, G., and Subramani, Chinnamuthu
- Subjects
MACHINE learning ,ROTATIONAL motion ,MAGNETIC fields ,POROUS materials ,MAGNETIC field effects ,SUPPORT vector machines ,GYROTRONS ,ARTIFICIAL neural networks - Abstract
Effects of the magnetic field and inertia on the onset of thermal convection in a horizontal bidispersive porous layer, rotating about a vertical axis, are analyzed. The Darcy equation with same temperature in the micro- and macrophases is used to characterize the fluid motion. The Vadasz number is taken into account in a generalized Darcy equation for the macrophase. The eigenvalue problem obtained from the linear stability analysis is solved analytically for free–free boundaries. Moving one step further from the traditional linear stability analysis, machine learning tools are introduced in this paper to include the effect of multiple parameters on the marginal state of the system. Machine learning techniques have been implemented to identify the mode of instability with respect to different parameters. In particular, classification algorithms, namely, Artificial Neural Networks (ANN) and Support vector machine, are used to examine the onset of oscillatory convection and stationary convection. The required data for training of the algorithms are generated from the results of linear stability analysis. It is found that ANN with the sufficient number of hidden layers along with good choice of training dataset can predict the mode of instability even on the small variation in a given parameter. The combined effect of rotation, magnetic field, and inertia is to reduce the oscillatory mode of instability; hence, the system exhibits the steady mode of instability for a significant region in the three dimensional space comprising the Taylor number, the Hartman number, and the Vadasz number. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
15. A systematic review for detecting cancer using machine learning techniques.
- Author
-
Sharma, Geetika and Prabha, Chander
- Subjects
ARTIFICIAL neural networks ,CATHODE ray tubes ,MACHINE theory ,RECEIVER operating characteristic curves ,BAYESIAN analysis ,MACHINE learning ,SUPPORT vector machines ,K-nearest neighbor classification - Abstract
Machine learning has helped a lot in the field of cancer research. For the detection and diagnosis of cancer various algorithms like artificial neural networks, Support Vector Machine, Decision Tree, Bayesian Networks, Naive Bayes and K-Nearest Neighbor etc. have been used. Today Machine learning techniques are widely used in distinguishing and characterizing diseases by means of X-ray and CRT (Cathode Ray Tube) images. Using these techniques, one is attempting to anticipate the probability of building up of cancer growth before the disease occur. In this paper, a review has been done on different approaches used to detect different type of cancer using various machine learning techniques. Further the efficiency of these methods is compared on the basis of various parameters like Accuracy, Sensitivity, Specificity, F-measure and ROC (Receiver Operating Characteristics) Curve. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
16. Machine learning based fraud detection in credit card data transactions.
- Author
-
Tamizharasi, A., Rose, S. Remya, Rao, K. Veerabhadra, Reddy, K. Mohith, and Varun, J. Krishna
- Subjects
CREDIT card fraud ,SUPERVISED learning ,ARTIFICIAL neural networks ,SWINDLERS & swindling ,INSTALLMENT plan ,TEXT messages ,MACHINE learning - Abstract
Improvement of correspondence advancements and web based business has made the Master card as them ost well -known method of installment for both on the web band standard buys. Along these lines, security during this framework is extremely expected to stop misrepresentation exchanges. Mis representation exchanges in Visa information exchange are expanding yearly. Toward this path, analysts additionally are attempting the novel strategies to distinguish hand stop such cheats. In any case, there is consistently a need of certain procedures that ought to unequivocally and productively identify these fakes. This article includes a plan for the purpose of con artists in the visa details, with the help of the neural network (NN), unsupervised learning in the process. The proposed method is superior to the common methods of Auto Encoder (AE), the Local Emission Factor (LOF), isolation forest (IF),and the K-means clustering. The proposed method for the localization of the expression is based on the NNR operates with an accuracy of 99.87%, compared to the existing, AE, AS, PRAISE, and 'K' Means as the strategies that will give an accuracy of about 97c/o,98c/o,98c/o, and 99.75c/o separately. Frauds in credit card transactions are common today as most of us are using the credit card payment methods more frequently. This is due to the advancement of Technology and increase in online transaction resulting in frauds causing huge financial loss. Therefore, there is need for effective methods to reduce the loss. In addition, fraudsters find ways to steal the credit card information of the user by sending fake SMS and calls, also through masquerading attack, phishing attack and so on. This paper aims in using the multiple algorithms of Machine learning such as support vector machine (SVM), k-nearest neighbor (Knn) and artificial neural network (ANN) in predicting the occurrence of the fraud. Further, we conduct a differentiation of the accomplished supervised machine learning and deep learning techniques to differentiate between fraud and non-fraud transactions. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
17. Voice analysis for personal identification using FFT, machine learning and AI techniques.
- Author
-
Balabanova, Ivelina, Georgiev, Georgi, Karapenev, Boyan, and Rankovska, Valentina
- Subjects
VOICE analysis ,ARTIFICIAL neural networks ,ARTIFICIAL intelligence ,MARQUARDT algorithm ,FEATURE extraction ,MACHINE learning - Abstract
In this paper a hybrid approach for spectral analysis and voice profiles recognition by techniques on the base of Machine Learning and Artificial Intelligence (AI) have been proposed. For voice-processing procedures was applied the algorithm of Fast Fourier Transformation (FFT) with several different window types, respectively Hamming, 4 Term BHarris, Flat Top and Hanning. The spectral feature extraction is used for pre-processing training sets about k-Nearest Neighbours (k-NN) classification models and Feed-Forward Neural Networks (FFNN) for individual's personal identity about target group of people. Euclidean, Cityblock, Minkowski and Chebychev metric distances were applied in k-NN model creation. The design models are evaluated through resubstitution and cross-validation techniques. Levenberg- Marquardt learning algorithm was used to FFNN architectures with Linear and Tangent Sigmoid activation functionsin network outputs. High quality k-NN and FFNN models in regard to personal voice identification with level of accuracy achieved 97.68 % and 100.0 % were synthesized. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
18. Recurrent networks for direction-of-arrival identification of an acoustic source in a shallow water channel using a vector sensora).
- Author
-
Whitaker, Steven, Barnard, Andrew, Anderson, George D., and Havens, Timothy C.
- Subjects
ACOUSTIC radiators ,ARTIFICIAL neural networks ,RECURRENT neural networks ,MACHINE learning ,WATER use ,ACOUSTIC transducers ,WATER depth ,ACOUSTIC emission testing - Abstract
Conventional direction-of-arrival (DOA) estimation algorithms for shallow water environments usually contain high amounts of error due to the presence of many acoustic reflective surfaces and scattering fields. Utilizing data from a single acoustic vector sensor, the magnitude and DOA of an acoustic signature can be estimated; as such, DOA algorithms are used to reduce the error in these estimations. Three experiments were conducted using a moving boat as an acoustic target in a waterway in Houghton, Michigan. The shallow and narrow waterway is a complex and non-linear environment for DOA estimation. This paper compares minimizing DOA errors using conventional and machine learning algorithms. The conventional algorithm uses frequency-masking averaging, and the machine learning algorithms incorporate two recurrent neural network architectures, one shallow and one deep network. Results show that the deep neural network models the shallow water environment better than the shallow neural network, and both networks are superior in performance to the frequency-masking average method. [ABSTRACT FROM AUTHOR]
- Published
- 2021
- Full Text
- View/download PDF
19. Mouse livers machine learning identification based on hyperspectral x-ray computed tomography reconstructed x-ray absorption spectra.
- Author
-
Fang, Zheng, Zhong, Shuo, Hu, Weifeng, and Cheng, Siyuan
- Subjects
X-ray absorption spectra ,COMPUTER-assisted image analysis (Medicine) ,MACHINE learning ,COMPUTED tomography ,ARTIFICIAL neural networks ,POSITRON emission - Abstract
X-ray computed tomography (X-CT) is often used to examine organs, but the reconstructed images can only be used for structural identification. Whether the organs are healthy or not requires a professional doctor to examine the reconstructed image and judge from his or her own experience. The purpose of this paper is to identify the cirrhotic mouse liver and normal mouse liver with hyperspectral x-ray CT (HXCT) and machine learning. HXCT is proposed to reconstruct the x-ray absorption spectrum (XAS) characteristics of a single pixel in the reconstructed mouse liver images. HXCT uses a cadmium telluride photon counter as the x-ray detector, which can improve the spectral resolution and separate spectral lines. Filtered back-projection and algebra reconstruction technique reconstruction algorithms are used for image and XAS reconstruction. In the machine learning model, principal component analysis is utilized to reduce the dimensionality of XAS. Besides, the neural network algorithm Artificial Neural Network (ANN) is used to train and identify the reconstructed XAS of two different kinds of livers. These two different mouse livers can be well recognized since the accuracy goes to almost 100% based on ANN. It is feasible to employ the machine learning algorithm to identify the XAS of different mouse livers. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
20. Application of Statistical Techniques in Environmental Modelling.
- Author
-
Yahaya, Ahmad Shukri
- Subjects
STATISTICAL accuracy ,MACHINE learning ,LEAST squares ,ARTIFICIAL neural networks ,SUPPORT vector machines ,DISTRIBUTION (Probability theory) ,REGRESSION trees - Abstract
The aim of statistical modeling is to predict with good accuracy the variable of interest and sometimes to determine the contribution of the independent variables towards the dependent variable. Over the years, researches have always tried to produce better models in order to increase the accuracy of their predictions. This paper will discuss some statistical techniques that were used frequently in environmental research as well as recent techniques and advances in statistical modelling. The first technique that will be discussed is the multiple linear regression (MLR) models. Some other regression models will also be given. Next probability distributions will be discussed. This is to obtain statistical distributions that best fit the variable of interest. This will enable the researcher to predict the return period of an event. The application of Bayesian statistics has increased over the years. The used of least squares method in MLR and hypothesis testing is known as classical statistical techniques. The Bayesian statistics is based on the probability distributions of the prior and likelihood of the Bayes theorem. The results of using classical and Bayesian statistics will be discussed. Machine learning method will be deliberated next. Machine learning is the application of artificial intelligence that provides systems the ability to automatically learn and improve from experience without being programmed explicitly. Some machine learning methods that will be discussed are boosted regression trees, artificial neural networks and support vector machines. Finally, hybrid models that are the combination of two statistical models will be deliberated next. It is hoped that by combining two statistical models, a more accurate and improved model will be obtained. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
21. Visual Surveillance for Collective Behavior Analysis: From Human to Fish.
- Author
-
Hitoshi Habe
- Subjects
COLLECTIVE behavior ,HUMAN beings ,ANIMAL ecology ,SOCIAL groups ,ARTIFICIAL neural networks ,FISH schooling ,FISH locomotion - Abstract
Animals including human being are often with others in groups such as people crowds, fish schools, and flocks of birds. In such groups, they are interacting with each other to tell their intentions. By analyzing such collective behavior, we can obtain plenty of information not only for understanding the ecology of animals but also for practical applications including aquaculture and social safety. Hence, collective behavior analysis has been an active research area of both ecology and engineering. In this paper, we will introduce our recent research projects for visual surveillance of collective behavior analysis. First, we will present small object detection based on a deep neural network. The method is specially designed for detecting small swimming fish. Although training deep neural network usually requires a large number of annotated data, our proposed method reduces the cost of manual annotation. Next, we will present social group detection based on multiple instance learning. The multiple instance learning enables us to extract meaningful information from given data. We examined both of the proposed method using actual data to show their performance. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
22. Application of machine learning to spectrum and image data.
- Author
-
Aoyagi, Satoka
- Subjects
MACHINE learning ,SECONDARY ion mass spectrometry ,ARTIFICIAL neural networks ,SUPERVISED learning ,OPEN learning ,DATA mining ,PRINCIPAL components analysis - Abstract
Machine learning is a useful tool when extracting hidden information from complex measurement data obtained via surface analysis, as in secondary ion mass spectrometry. Flexible learning methods often require significant effort to adjust parameters, as these parameters may have a significant effect on results. However, machine learning methods enable the extraction of new information that cannot be found by manual analysis. This paper presents some examples of complex data analyses using conventional multivariate analysis methods based on linear combinations (principal component analysis and multivariate curve resolution), an unsupervised learning method based on artificial neural networks (sparse autoencoder), and a supervised learning method based on decision trees (random forest). To obtain reproducible and useful results from machine learning applications to surface analysis data, the preparation of data sets—including the selection of variables and the raw data conversion process—is crucial. Moreover, sufficient information representing analytical purposes, such as the chemical structures of unknown samples, material types, and physical or chemical properties of particular materials, must be contained in the data set for supervised learning. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
23. Seeing double with a multifunctional reservoir computer.
- Author
-
Flynn, Andrew, Tsachouridis, Vassilios A., and Amann, Andreas
- Subjects
- *
ARTIFICIAL neural networks , *BIOLOGICAL neural networks , *COMPUTERS , *DYNAMICAL systems , *MACHINE learning - Abstract
Multifunctional biological neural networks exploit multistability in order to perform multiple tasks without changing any network properties. Enabling artificial neural networks (ANNs) to obtain certain multistabilities in order to perform several tasks, where each task is related to a particular attractor in the network's state space, naturally has many benefits from a machine learning perspective. Given the association to multistability, in this paper, we explore how the relationship between different attractors influences the ability of a reservoir computer (RC), which is a dynamical system in the form of an ANN, to achieve multifunctionality. We construct the "seeing double" problem in order to systematically study how a RC reconstructs a coexistence of attractors when there is an overlap between them. As the amount of overlap increases, we discover that for multifunctionality to occur, there is a critical dependence on a suitable choice of the spectral radius for the RC's internal network connections. A bifurcation analysis reveals how multifunctionality emerges and is destroyed as the RC enters a chaotic regime that can lead to chaotic itinerancy. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
24. Modeling and Control with Neural Networks through a New Learning Method.
- Author
-
Nagy, Endre
- Subjects
ARTIFICIAL neural networks ,AUTOMATIC control systems ,MACHINE learning ,AUTOMATIC identification ,PARAMETER estimation - Abstract
New learning methods are presented in the paper to train neurons in a neural network. The methods are based on the principle "successive group optimization / identification", which states that parameters of a feedforward neural network may be estimated through layer by layer successive optimizations and iteration. Estimations on the parameters may be achieved in different ways; two possibilities are shown in the paper. Solutions of basic control problems with the proposed learning method and previously developed methods are also dealt with. [ABSTRACT FROM AUTHOR]
- Published
- 2014
- Full Text
- View/download PDF
25. Application of machine learning methods in photovoltaic output power prediction: A review.
- Author
-
Zhang, Wenyong, Li, Qingwei, and He, Qifeng
- Subjects
MACHINE learning ,SUPPORT vector machines ,ARTIFICIAL neural networks ,BLENDED learning ,PREDICTION models ,ENERGY consumption ,FORECASTING - Abstract
As the proportion of photovoltaic (PV) power generation rapidly increases, accurate PV output power prediction becomes more crucial to energy efficiency and renewable energy production. There are numerous approaches for PV output power prediction. Many researchers have previously summarized PV output power prediction from different angles. However, there are relatively few studies that use machine learning methods as a means to conduct a separate review of PV output power prediction. This review classifies machine learning methods from different perspectives and provides a systematic and critical review of machine learning methods for recent PV output power applications in terms of the temporal and spatial scales of prediction and finds that the artificial neural network and support vector machine are used much more frequently than other methods. In addition, this study examines the differences between the output power prediction of individual PV plants and regional PV stations and the benefits of regional PV plant prediction, while this paper presents some performance evaluation matrices commonly used for PV output power prediction. In addition, to further improve the accuracy of machine learning methods for PV output power prediction, some researchers suggest preprocessing the input data of the prediction models or considering hybrid machine learning methods. Furthermore, the potential advantages of machine model optimization for prediction performance improvement are discussed and explored in detail. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
26. Predictions of Reynolds and Nusselt numbers in turbulent convection using machine-learning models.
- Author
-
Bhattacharya, Shashwat, Verma, Mahendra K., and Bhattacharya, Arnab
- Subjects
NUSSELT number ,REYNOLDS number ,MACHINE learning ,ARTIFICIAL neural networks ,RAYLEIGH number ,FORECASTING - Abstract
In this paper, we develop a multivariate regression model and a neural network model to predict the Reynolds number (Re) and Nusselt number in turbulent thermal convection. We compare their predictions with those of earlier models of convection: Grossmann–Lohse [Phys. Rev. Lett. 86, 3316 (2001)], revised Grossmann–Lohse [Phys. Fluids 33, 015113 (2021)], and Pandey–Verma [Phys. Rev. E 94, 053106 (2016)] models. We observe that although the predictions of all the models are quite close to each other, the machine-learning models developed in this work provide the best match with the experimental and numerical results. [ABSTRACT FROM AUTHOR]
- Published
- 2022
- Full Text
- View/download PDF
27. X - ray Weld Image Classification Using Improved Convolutional Neural Network.
- Author
-
Nana Yang, Haijun Niu, Liang Chen, and Guihua Mi
- Subjects
X-rays ,ARTIFICIAL neural networks ,ARTIFICIAL intelligence ,ELECTROMAGNETIC waves ,MACHINE learning - Abstract
The defect detection of X-ray weld images is an effective method to improve product quality and safety. Due to the low contrast of the image, the way of using traditional feature extraction and machine learning has low accuracy. In this paper, combined theory with practice, the technique based on improved convolutional neural network is proposed to classify X-ray weld images. In comparison with conventional methods, it avoids de-noising, extracting and enhancing features. Experimental results on the images obtained from the actual production show that the introduced method has superior accuracy of the classification. [ABSTRACT FROM AUTHOR]
- Published
- 2018
- Full Text
- View/download PDF
28. An overload behavior detection system for engineering transport vehicles based on deep learning.
- Author
-
Zhou, Libo, Wu, Gang, Liu, Lin, Yang, Can, and Ke, Jianfeng
- Subjects
TRANSPORT vehicles ,ARTIFICIAL neural networks ,DEEP learning ,MACHINE learning ,DETECTORS - Abstract
This paper builds an overloaded truck detect system called ITMD to help traffic department automatically identify the engineering transport vehicles (commonly known as ‘dirt truck’) in CCTV and determine whether the truck is overloaded or not. We build the ITMD system based on the Single Shot MultiBox Detector (SSD) model. By constructing the image dataset of the truck and adjusting hyper-parameters of the original SSD neural network, we successfully trained a basic network model which the ITMD system depends on. The basic ITMD system achieves 83.01% mAP on classifying overload/non-overload truck, which is a not bad result. Still, some shortcomings of basic ITMD system have been targeted to enhance: it is easy for the ITMD system to misclassify other similar vehicle as truck. In response to this problem, we optimized the basic ITMD system, which effectively reduced basic model’s false recognition rate. The optimized ITMD system achieved 86.18% mAP on the test set, which is better than the 83.01% mAP of the basic ITMD system. [ABSTRACT FROM AUTHOR]
- Published
- 2018
- Full Text
- View/download PDF
29. Literature review of breast cancer detection using machine learning algorithms.
- Author
-
Abunasser, Basem S., AL-Hiealy, Mohammed Rasheed J., Zaqout, Ihab S., and Abu-Naser, Samy S.
- Subjects
MACHINE learning ,DEEP learning ,ARTIFICIAL neural networks ,EARLY detection of cancer ,BREAST cancer - Abstract
Cancer is the leading cause of non-accidental deaths worldwide. Specifically, nearly 10 million people died globally from cancer in the year 2020. Breast Cancer (BC) is a common and fatal disease among women worldwide, and ranks fourth among the fatal diseases among various cancers, such as cervical, colorectal, and cervical tumors and brain tumors. In addition, the number of new breast cancer patients is expected to increase by 70% in the next 20 years. Therefore, early and accurate diagnosis plays a pivotal role in improving prognosis and increasing the survival rate of cancer patients from 30 to 50%. With technical advances in healthcare, machine learning and deep learning play an important role in processing and analyzing a large number of medical images. The aim of this study is to identify studies that have been done on the application of classification techniques in diagnosing BC and analyze them from four perspectives: classification techniques used, Dataset used, Programming language used and best accuracy. We conducted a systematic literature review of 32 selected studies published between 2002 and 2020. The results showed that among the classification techniques examined, artificial neural networks, support vector machines and k-nearest neighbor were the most widely used. Moreover, artificial neural networks, support vector machines, and group classifiers have been implemented better than other techniques, with average accuracy values between 83.45% and 99.30%. Most of the selected studies used Wisconsin CSV dataset and a few of the studies used different types of images such as mammography, ultrasound, and micro images. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
30. Low-Cost Autonomous Perceptron Neural Network Inspired by Quantum Computation.
- Author
-
Zidan, Mohammed, Abdel-Aty, Abdel-Haleem, El-Sadek, Alaa, Zanaty, E. A., and Abdel-Aty, Mahmoud
- Subjects
ARTIFICIAL neural networks ,QUANTUM computing ,CONSCIOUS automata ,QUANTUM mechanics ,MACHINE learning - Abstract
Achieving low cost learning with reliable accuracy is one of the important goals to achieve intelligent machines to save time, energy and perform learning process over limited computational resources machines. In this paper, we propose an efficient algorithm for a perceptron neural network inspired by quantum computing composite from a single neuron to classify inspirable linear applications after a single training iteration O(1). The algorithm is applied over a real world data set and the results are outer performs the other state-of-the art algorithms. [ABSTRACT FROM AUTHOR]
- Published
- 2017
- Full Text
- View/download PDF
31. Research on AHP Decision Algorithms Based on BP Algorithm.
- Author
-
Ning Ma and Jianhe Guan
- Subjects
ANALYTIC hierarchy process ,DECISION making in science ,ARTIFICIAL neural networks ,NONLINEAR systems ,MACHINE learning - Abstract
Decision making is the thinking activity that people choose or judge, and scientific decision-making has always been a hot issue in the field of research. Analytic Hierarchy Process (AHP) is a simple and practical multi-criteria and multi-objective decision-making method that combines quantitative and qualitative and can show and calculate the subjective judgment in digital form. In the process of decision analysis using AHP method, the rationality of the twodimensional judgment matrix has a great influence on the decision result. However, in dealing with the real problem, the judgment matrix produced by the two-dimensional comparison is often inconsistent, that is, it does not meet the consistency requirements. BP neural network algorithm is an adaptive nonlinear dynamic system. It has powerful collective computing ability and learning ability. It can perfect the data by constantly modifying the weights and thresholds of the network to achieve the goal of minimizing the mean square error. In this paper, the BP algorithm is used to deal with the consistency of the two-dimensional judgment matrix of the AHP. [ABSTRACT FROM AUTHOR]
- Published
- 2017
- Full Text
- View/download PDF
32. On the Fusion of Tuning Parameters of Fuzzy Rules and Neural Network.
- Author
-
Mamuda, Mamman and Sathasivam, Saratha
- Subjects
FUZZY logic ,ARTIFICIAL neural networks ,MACHINE learning ,RADIAL basis functions - Abstract
Learning fuzzy rule-based system with neural network can lead to a precise valuable empathy of several problems. Fuzzy logic offers a simple way to reach at a definite conclusion based upon its vague, ambiguous, imprecise, noisy or missing input information. Conventional learning algorithm for tuning parameters of fuzzy rules using training input-output data usually end in a weak firing state, this certainly powers the fuzzy rule and makes it insecure for a multiple-input fuzzy system. In this paper, we introduce a new learning algorithm for tuning the parameters of the fuzzy rules alongside with radial basis function neural network (RBFNN) in training input-output data based on the gradient descent method. By the new learning algorithm, the problem of weak firing using the conventional method was addressed. We illustrated the efficiency of our new learning algorithm by means of numerical examples. MATLAB R2014(a) software was used in simulating our result The result shows that the new learning method has the best advantage of training the fuzzy rules without tempering with the fuzzy rule table which allowed a membership function of the rule to be used more than one time in the fuzzy rule base. [ABSTRACT FROM AUTHOR]
- Published
- 2017
- Full Text
- View/download PDF
33. Online Transfer Learning with Extreme Learning Machine.
- Author
-
Haibo Yin and Yun-an Yang
- Subjects
DEEP learning ,MACHINE learning ,KNOWLEDGE representation (Information theory) ,ARTIFICIAL neural networks ,DISTANCE education - Abstract
In this paper, we propose a new transfer learning algorithm for online training. The proposed algorithm, which is called Online Transfer Extreme Learning Machine (OTELM), is based on Online Sequential Extreme Learning Machine (OSELM) while it introduces Semi-Supervised Extreme Learning Machine (SSELM) to transfer knowledge from the source to the target domain. With the manifold regularization, SSELM picks out instances from the source domain that are less relevant to those in the target domain to initialize the online training, so as to improve the classification performance. Experimental results demonstrate that the proposed OTELM can effectively use instances in the source domain to enhance the learning performance. [ABSTRACT FROM AUTHOR]
- Published
- 2017
- Full Text
- View/download PDF
34. Predictive modelling of the penetration coefficient of cold metal transfer welded joints using machine learning approaches.
- Author
-
Kumar, Nalajam Pavan, Varadarajan, Ramesh, Gupta, M Satyanarayana, Gupta, TVK, and Nath, N Kishore
- Subjects
MACHINE learning ,WELDED joints ,KRIGING ,ARTIFICIAL neural networks ,ROBOTIC welding - Abstract
Weld bead geometry is an important characteristic of the weld joints in evaluating its quality. However, weld geometry is characterized by three parameters namely weld width, weld height and weld depth. The measurement system of three parameters is often time consuming, particularly for intelligent robots used in the welding processes. Hence, in this paper, we introduced a penetration coefficient (PC) that effectively minimizes the complexities involved in existing measurement system of weld geometry. Further, various machine learning approaches are used to predict the penetration coefficient. Cold metal transfer welded AA6061 sheets are chosen to obtain the data of penetration coefficient. Linear regression (LR), support vector machine (SVM) regression and Gaussian process regression (GPR) models and artificial neural network (ANN) model are used for predictive modelling. The statistical performance factors of the models reveal the superior performance of ANN model. The lowest mean absolute error of 0.15 is observed for ANN followed by the SVM (0.31), GPR (0.39) and LR (0.41). [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
35. Machine learning techniques for fault detection of induction motor bearings.
- Author
-
Elango, Murugappan, Annamalai, Adithyan, Venkateswaran, A., Sasipraba, T, Subramaniam, Prakash, Jayaprabakar, J, Joy, Nivin, Anish, M, Ganesan, S, and Kavitha, K R
- Subjects
BEARINGS (Machinery) ,INDUCTION machinery ,SYSTEM downtime ,MACHINE learning ,ELECTRIC motors ,ARTIFICIAL neural networks ,ELECTRIC drives - Abstract
This paper gives an approach based on a fitting technique to detect faulty bearings by employing artificial neural networks (ANNs). At present, many companies follow very expensive, scheduled maintenance procedures for checking the bearing condition of the electric drives. Therefore, there is considerable demand to reduce the maintenance costs and to prevent unscheduled downtime of electrical drives in company. To overcome the above crisis, online monitoring of vibrations in electric motors at running condition can be implemented. Most of the vibration measuring equipment measure the total vibration of the electric motor. Deducing the bearing fault from the total vibration value of the motor is not accurate. Hence it is quintessential to split the total vibration value of the motor into its sub-components namely bearing vibration and the vibration due to loading. This splitting could visualize the exact occurrence of fault in the machine. For the known loading conditions of the motor, the total vibration, bearing vibration and the vibration due to loading are tabulated. Some tabulated data are used to train the ANNs for the known machine conditions. Once trained, the ANNs are examined using the remaining data. The final evaluated ANNs can readily split the total vibration value into its sub-components enabling the computation of component value. These sub-component vibration values are analyzed individually for abnormalities to find the bearing fault accurately. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
36. Enhanced sparse matrix approach in neural network algorithm for an effective intelligent classification system.
- Author
-
Sagir, Abdu Masanawa, Abubakar, Hamza, Ibrahim, Siti Nur Iqmal, Ibrahim, Noor Akma, Ismail, Fudziah, Lee, Lai Soon, Leong, Wah June, Midi, Habshah, and Wahi, Nadihah
- Subjects
SPARSE matrices ,ALGORITHMS ,MACHINE learning ,CLASSIFICATION ,KEY performance indicators (Management) ,ARTIFICIAL neural networks - Abstract
The main objective of this research is to develop an effective intelligent system that can be used by medical practitioners (physicians) to accelerate diagnosis and treatment processes. In this paper, the sparse matrix approach was incorporated in neural network learning algorithm for scalability, minimize higher memory usage/storage capacity, enhancing implementation time and speed up the analysis of the data. The proposed intelligent classification system maximizes the intelligently classification of data and minimizes the number of trends inaccurately identified. For robustness, the proposed method was tested with three different datasets, namely, Hepatitis, SPECT Heart and Cleveland Heart. Therefore, an attempt was made to determine the performance indicators efficacy. Compared to some similar existing methods, the approach presented achieves improved performance. The program used for implementation of the proposed model is MATLAB R2016a (version 9.0) and executed in the 4.0 GB RAM processor of PC Intel Pentium Quad Core N3700. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
37. Moisture-Loss Prediction System in Withering of Pepper using Machine Learning.
- Author
-
Aishwarya, G. and Raj, Dhivya
- Subjects
MACHINE learning ,FORECASTING ,PEPPERS ,MANUFACTURING processes ,INTELLIGENT sensors ,HUMIDITY ,ARTIFICIAL neural networks - Abstract
Spices and beverages plays an important role in every cuisine. The quality of these products should be good enough for their best use. The processing steps involved during their manufacturing and the implementation methodology determines the quality of these products. Among spices, Pepper is known as the king of spices. For assuring the quality of pepper, better methods during processing must be adopted. Withering is the most crucial process in the manufacturing of black pepper. It determines the quality and durability of the final product. In this paper, a system is implemented to attain the desired Moisture Loss (ML) during the withering process of black pepper. This is achieved by predicting in advance the ML using Machine Learning algorithm. The main factors influencing the prediction process are inlet and outlet temperature and relative humidity. A prototype trough is developed with smart sensor nodes placed at the inlet and outlet for measuring moisture and temperature. The data measured is saved in a database and this is utilised to predict the ML using ANN(Artificial Neural Network). The error between predictedML and actual ML due to weight loss is analysed. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
38. Predicting Dynamic Behavior of a Biological System Using ANNs.
- Author
-
Osman, Mohd Haniff, Ibrahim, Ratnawati, Hashim, Ishak, Liong Choong Yeun, Bakar, Azuraliza Abu, and Hussein, Zeti Azura Mohamed
- Subjects
BIOLOGICAL neural networks ,COGNITIVE neuroscience ,NEUROBIOLOGY ,NEURAL circuitry ,MEDICAL technology ,COGNITIVE science - Abstract
In this paper, artificial neural networks (ANNs) are applied to predict protein concentrations of a biological system. The input data are generated from a nonlinear mathematical model of the protein concentration. The protein concentrations from CDC6 data with actual kinetic parameter are taken as the target output. The data are then trained using multilayer perceptron (MLP) neural network with a 6-6-6 configuration. The allocation of the data will be distributed into 3 categories that are 80% as training data, 10% as validation data, and 10% as test data. The learning rules used in this work to determine the best model are gradient descent, conjugate gradient, scaled conjugate gradient. It is found that the MLP with scaled conjugate gradient learning rule gives the best prediction rate. [ABSTRACT FROM AUTHOR]
- Published
- 2008
- Full Text
- View/download PDF
39. Advancing flame retardant prediction: A self-enforcing machine learning approach for small datasets.
- Author
-
Yan, Cheng, Lin, Xiang, Feng, Xiaming, Yang, Hongyu, Mensah, Patrick, and Li, Guoqiang
- Subjects
ARTIFICIAL neural networks ,FIREPROOFING agents ,MACHINE learning ,FIRE resistant polymers ,FIREPROOFING ,HEAT release rates ,SPACE colonies ,ENTHALPY - Abstract
Improving the fireproof performance of polymers is crucial for ensuring human safety and enabling future space colonization. However, the complexity of the mechanisms for flame retardant and the need for customized material design pose significant challenges. To address these issues, we propose a machine learning (ML) framework based on substructure fingerprinting and self-enforcing deep neural networks (SDNN) to predict the fireproof performance of flame-retardant epoxy resins. Our model is based on a comprehensive understanding of the physical mechanisms of materials and can predict fireproof performance and eliminate the needs for properties descriptors, making it more convenient than previous ML models. With a dataset of only 163 samples, our SDNN models show an average prediction error of 3% for the limited oxygen index (LOI). They also provide satisfactory predictions for the peak of heat release rate PHR and total heat release (THR), with coefficient of determination (R
2 ) values of 0.87 and 0.85, respectively, and average prediction errors less than 17%. Our model outperforms the support vector model SVM for all three indices, making it a state-of-the-art study in the field of flame retardancy. We believe that our framework will be a valuable tool for the design and virtual screening of flame retardants and will contribute to the development of safer and more efficient polymer materials. [ABSTRACT FROM AUTHOR]- Published
- 2023
- Full Text
- View/download PDF
40. Design of coaxial coils using hybrid machine learning.
- Author
-
Chen, Jun, Wu, Zeliang, Bao, Guzhi, Chen, L. Q., and Zhang, Weiping
- Subjects
BLENDED learning ,MACHINE learning ,ARTIFICIAL neural networks ,DIFFERENTIAL evolution ,COAXIAL cables ,MAGNETIC shielding ,MACHINING - Abstract
A coil system to generate a uniform field is urgently needed in quantum experiments. However, general coil configurations based on the analytical method have not considered practical restrictions, such as the region for coil placement due to holes in the center of the magnetic shield, which could not be directly applied in most of the quantum experiments. In this paper, we develop a coil design method for quantum experiments using hybrid machine learning. The algorithm part consists of a machine learner based on an artificial neural network and a differential evolution (DE) learner. The cooperation of both learners demonstrates its higher efficiency than a single DE learner and robustness in the coil optimization problem compared with analytical proposals. With the help of a DE learner, in numerical simulation, a machine learner can successfully design coaxial coil systems that generate fields whose relative inhomogeneity in a 25 mm-long central region is ∼10
−6 under constraints. In addition, for experiments, a coil system with 0.069% inhomogeneity of the field, designed by a machine learner, is constructed, which is mainly limited by machining the precision of the circuit board. Benefitting from machine learning's high-dimension optimization capabilities, our coil design method is convenient and has potential for various quantum experiments. [ABSTRACT FROM AUTHOR]- Published
- 2021
- Full Text
- View/download PDF
41. Implement data mining and deep learning techniques to detect financial distress.
- Author
-
Abdullah, Dalya Abdulkarim and AL-Anber, Nashaat Jasim
- Subjects
- *
DEEP learning , *DATA mining , *MACHINE learning , *NAIVE Bayes classification , *ARTIFICIAL neural networks , *BUSINESS enterprises - Abstract
Financial markets are currently a key source of growth for the local and international economies, as they are the means by which economic units are fed. Iraq is one of the countries attempting to improve and modernize its financial industry in order to stay up with technology advancements and the digital revolution. Forecasting and early detection of financial distress is one of the important methods that contribute to increasing confidence between investors and the market and help to make sound decisions in a timely manner in order to avoid reaching bankruptcy. This paper aims to employ smart models in the detection of financial distress, and to select the best model capable of classifying the financial situation of companies into three categories (non-distress, medium distress and high distress) by selecting (14) financial ratio that directly affects the situation of companies. The researcher used artificial neural networks algorithms such as the reverse error propagation algorithm, Using data mining methods and deep learning algorithms, the C4.5 algorithm, Naive Bayes simple classifier algorithm, Convolutional neural networks algorithm, and a multi-layered algorithm support vector machine were used to classify a company's financial state. The C4.5 method and SVM had the highest rating accuracy by a tiny margin in all levels (98, 96.9, and 91.9) respectively, according to the results of the analysis. The most essential recommendations included the fundamental requirement of using smart technology in recognizing financial challenges of companies in order to support and consolidate the economic stability of enterprises in particular and the market in general in the adoption of the Iraqi stock market. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
42. Modeling of sub-grid conditional mixing statistics in turbulent sprays using machine learning methods.
- Author
-
Yao, S., Wang, B., Kronenburg, A., and Stein, O. T.
- Subjects
MACHINE learning ,SPRAY combustion ,ARTIFICIAL neural networks ,COMBUSTION gases ,TURBULENT mixing ,FREE-space optical technology - Abstract
Deep artificial neural networks (ANNs) are used for modeling sub-grid scale mixing quantities such as the filtered density function (FDF) of the mixture fraction and the conditional scalar dissipation rate. A deep ANN with four hidden layers is trained with carrier-phase direct numerical simulations (CP-DNS) of turbulent spray combustion. A priori validation corroborates that ANN predictions of the mixture fraction FDF and the conditional scalar dissipation rate are in very good agreement with CP-DNS data. ANN modeled solutions show much better performance with a mean error of around 1%, which is one order of magnitude smaller than that of standard modeling approaches such as the β-FDF and its modified version. The predicted conditional scalar dissipation rate agrees very well with CP-DNS data over the entire mixture fraction space, whereas conventional models derived for pure gas phase combustion fail to describe ⟨N|ξ = η⟩ in regions with higher mixture fraction and low probability. In the second part of this paper, uncertainties associated with ANN predictions are analyzed. It is shown that a suitable selection of training sets can reduce the size of the necessary test database by ∼50% without compromising the accuracy. Feature importance analysis is used to analyze the importance of different combustion model parameters. While the droplet evaporating rate, the droplet number density, and the mixture fraction remain the dominant features, the influence of turbulence related parameters only becomes important if turbulence levels are sufficiently high. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
43. Machine learning surrogate models for Landau fluid closure.
- Author
-
Ma, Chenhao, Zhu, Ben, Xu, Xue-Qiao, and Wang, Weixing
- Subjects
MULTILAYER perceptrons ,ARTIFICIAL neural networks ,DISCRETE Fourier transforms ,MACHINE learning ,CONFIGURATION space ,DEEP learning - Abstract
The first result of applying the machine/deep learning technique to the fluid closure problem is presented in this paper. As a start, three different types of neural networks [multilayer perceptron (MLP), convolutional neural network (CNN), and two-layer discrete Fourier transform (DFT) network] were constructed and trained to learn the well-known Hammett–Perkins Landau fluid closure in configuration space. We find that in order to train a well-preformed network, a minimum size of the training data set is needed; MLP also requires a minimum number of neurons in the hidden layers that equals the degrees of freedom in Fourier space, despite the fact that training data are being fed into the configuration space. Out of the three models, DFT performs the best for the clean data, most likely due to the existence of the simple Fourier expression for the Hammett–Perkins closure, but it is the least robust with respect to input noise. Overall, with appropriate tuning and optimization, all three neural networks are able to accurately predict the Hammett–Perkins closure and reproduce the intrinsic nonlocal feature, suggesting a promising path to calculating more sophisticated closures with the machine/deep learning technique. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
44. Coarse-scale PDEs from fine-scale observations via machine learning.
- Author
-
Lee, Seungjoon, Kooshkbaghi, Mahdi, Spiliotis, Konstantinos, Siettos, Constantinos I., and Kevrekidis, Ioannis G.
- Subjects
MACHINE learning ,GAUSSIAN processes ,PARTIAL differential equations ,SPACETIME ,ARTIFICIAL neural networks ,MODEL-driven software architecture - Abstract
Complex spatiotemporal dynamics of physicochemical processes are often modeled at a microscopic level (through, e.g., atomistic, agent-based, or lattice models) based on first principles. Some of these processes can also be successfully modeled at the macroscopic level using, e.g., partial differential equations (PDEs) describing the evolution of the right few macroscopic observables (e.g., concentration and momentum fields). Deriving good macroscopic descriptions (the so-called "closure problem") is often a time-consuming process requiring deep understanding/intuition about the system of interest. Recent developments in data science provide alternative ways to effectively extract/learn accurate macroscopic descriptions approximating the underlying microscopic observations. In this paper, we introduce a data-driven framework for the identification of unavailable coarse-scale PDEs from microscopic observations via machine-learning algorithms. Specifically, using Gaussian processes, artificial neural networks, and/or diffusion maps, the proposed framework uncovers the relation between the relevant macroscopic space fields and their time evolution (the right-hand side of the explicitly unavailable macroscopic PDE). Interestingly, several choices equally representative of the data can be discovered. The framework will be illustrated through the data-driven discovery of macroscopic, concentration-level PDEs resulting from a fine-scale, lattice Boltzmann level model of a reaction/transport process. Once the coarse evolution law is identified, it can be simulated to produce long-term macroscopic predictions. Different features (pros as well as cons) of alternative machine-learning algorithms for performing this task (Gaussian processes and artificial neural networks) are presented and discussed. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF
45. Network physiology in insomnia patients: Assessment of relevant changes in network topology with interpretable machine learning models.
- Author
-
Jansen, Christoph, Penzel, Thomas, Hodel, Stephan, Breuer, Stefanie, Spott, Martin, and Krefting, Dagmar
- Subjects
INSOMNIACS ,BIG data ,MACHINE learning ,ARTIFICIAL neural networks ,SLEEP stages ,DIGITAL image processing - Abstract
Network physiology describes the human body as a complex network of interacting organ systems. It has been applied successfully to determine topological changes in different sleep stages. However, the number of network links can quickly grow above the number of parameters that are typically analyzed with standard statistical methods. Artificial Neural Networks (ANNs) are a promising approach as they are successful in large parameter spaces, such as in digital imaging. On the other hand, ANN models do not provide an intrinsic approach to interpret their predictions, and they typically require large training data sets. Both aspects are critical in biomedical research. Medical decisions need to be explainable, and large data sets of quality assured patient and control data are rare. In this paper, different models for the classification of insomnia—a common sleep disorder—have been trained with 59 patients and age and gender matched controls, based on their physiological networks. Feature relevance evaluation is employed for all methods. For ANNs, the extrinsic interpretation method DeepLift is applied. The results are not identical across methods, but certain network links have been rated as relevant by all or most of the models. While ANNs show less classification accuracy (0.89) than advanced tree-based models (0.92 and 0.93), DeepLift provides an in-depth ANN interpretation with feature relevance scores for individual data samples. The analysis revealed modifications in the pulmonar, ocular, and cerebral subnetworks that have not been described before but are consistent with known findings on the physiological impact of insomnia. [ABSTRACT FROM AUTHOR]
- Published
- 2019
- Full Text
- View/download PDF
46. Learned mappings for targeted free energy perturbation between peptide conformations.
- Author
-
Willow, Soohaeng Yoo, Kang, Lulu, and Minh, David D. L.
- Subjects
ARTIFICIAL neural networks ,PEPTIDES ,CONFIGURATION space ,OPTIMAL stopping (Mathematical statistics) ,MACHINE learning - Abstract
Targeted free energy perturbation uses an invertible mapping to promote configuration space overlap and the convergence of free energy estimates. However, developing suitable mappings can be challenging. Wirnsberger et al. [J. Chem. Phys. 153, 144112 (2020)] demonstrated the use of machine learning to train deep neural networks that map between Boltzmann distributions for different thermodynamic states. Here, we adapt their approach to the free energy differences of a flexible bonded molecule, deca-alanine, with harmonic biases and different spring centers. When the neural network is trained until "early stopping"—when the loss value of the test set increases—we calculate accurate free energy differences between thermodynamic states with spring centers separated by 1 Å and sometimes 2 Å. For more distant thermodynamic states, the mapping does not produce structures representative of the target state, and the method does not reproduce reference calculations. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
47. Chromium Distribution Forecasting Using Multilayer Perceptron Neural Network and Multilayer Perceptron Residual Kriging.
- Author
-
Tarasov, Dmitry, Buevich, Alexander, Shichkin, Andrey, Subbotina, Irina, Tyagunov, Andrey, and Baglaeva, Elena
- Subjects
CHROMIUM content of soils ,ARTIFICIAL neural networks ,MULTILAYER perceptrons ,GEOLOGICAL statistics ,MACHINE learning ,KRIGING - Abstract
It is known that combination of geostatistical interpolation techniques (e.g. kriging) and machine learning (e.g. neural networks) leads to better prediction accuracy and productivity. The paper deals with application of the artificial neural network residual kriging (ANNRK) to the spatial prediction of soil pollution by Chromium (Cr). In the work, we examined and compared two neural networks: Multilayer Perceptron (MLP) and Multilayer Perceptron Residual Kriging (MLPRK). The case study is based on the survey on surface contamination by Cr at the subarctic Noyabrsk, Russia. The proposed models have been built, implemented and validated using ArcGIS and MATLAB software. The models frameworks have been developed using a computer simulation based on a minimization of the root mean squared error (RMSE). Both models showed almost identical results. [ABSTRACT FROM AUTHOR]
- Published
- 2018
- Full Text
- View/download PDF
48. Development and application of deep convolutional neural network in target detection.
- Author
-
Jiang, Xiaowei, Wang, Chunping, Fu, Qiang, Liu, Lin, Yang, Can, and Ke, Jianfeng
- Subjects
BIG data ,ARTIFICIAL neural networks ,OPTICAL neural nets ,MACHINE learning ,MATHEMATICAL convolutions - Abstract
With the development of big data and algorithms, deep convolution neural networks with more hidden layers have more powerful feature learning and feature expression ability than traditional machine learning methods, making artificial intelligence surpass human level in many fields. This paper first reviews the development and application of deep convolutional neural networks in the field of object detection in recent years, then briefly summarizes and ponders some existing problems in the current research, and the future development of deep convolutional neural network is prospected. [ABSTRACT FROM AUTHOR]
- Published
- 2018
- Full Text
- View/download PDF
49. Boltzmann's equation at 150: Traditional and modern solution techniques for charged particles in neutral gases.
- Author
-
Boyle, G. J., Stokes, P. W., Robson, R. E., and White, R. D.
- Subjects
BOLTZMANN'S equation ,ELECTRIC discharges ,ARTIFICIAL neural networks ,GLOW discharges ,ELECTRON gas ,MACHINE learning - Abstract
Seminal gas discharge experiments of the late 19th and early 20th centuries laid the foundations of modern physics, and the influence of this "golden era" continues to resonate well into the 21st century through modern technologies, medical applications, and fundamental scientific investigations. Key to this continuing success story has been the kinetic equation formulated by Ludwig Boltzmann in 1872, which provides the theoretical foundations necessary for analyzing such highly non-equilibrium situations. However, as discussed here, the full potential of Boltzmann's equation has been realized only in the past 50 years or so, with modern computing power and analytical techniques facilitating accurate solutions for various types of charged particles (ions, electrons, positrons, and muons) in gases. Our example of thermalization of electrons in xenon gas highlights the need for such accurate methods—the traditional Lorentz approximation is shown to be hopelessly inadequate. We then discuss the emerging role of Boltzmann's equation in determining cross sections by inverting measured swarm experiment transport coefficient data using machine learning with artificial neural networks. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
50. Predictive understanding of the surface tension and velocity of sound in ionic liquids using machine learning.
- Author
-
Mohan, Mood, Smith, Micholas Dean, Demerdash, Omar, Kidder, Michelle K., and Smith, Jeremy C.
- Subjects
SURFACE tension ,MACHINE learning ,ARTIFICIAL neural networks ,IONIC liquids ,SPEED of sound - Abstract
Knowledge of the physical properties of ionic liquids (ILs), such as the surface tension and speed of sound, is important for both industrial and research applications. Unfortunately, technical challenges and costs limit exhaustive experimental screening efforts of ILs for these critical properties. Previous work has demonstrated that the use of quantum-mechanics-based thermochemical property prediction tools, such as the conductor-like screening model for real solvents, when combined with machine learning (ML) approaches, may provide an alternative pathway to guide the rapid screening and design of ILs for desired physiochemical properties. However, the question of which machine-learning approaches are most appropriate remains. In the present study, we examine how different ML architectures, ranging from tree-based approaches to feed-forward artificial neural networks, perform in generating nonlinear multivariate quantitative structure–property relationship models for the prediction of the temperature- and pressure-dependent surface tension of and speed of sound in ILs over a wide range of surface tensions (16.9–76.2 mN/m) and speeds of sound (1009.7–1992 m/s). The ML models are further interrogated using the powerful interpretation method, shapley additive explanations. We find that several different ML models provide high accuracy, according to traditional statistical metrics. The decision tree-based approaches appear to be the most accurate and precise, with extreme gradient-boosting trees and gradient-boosting trees being the best performers. However, our results also indicate that the promise of using machine-learning to gain deep insights into the underlying physics driving structure–property relationships in ILs may still be somewhat premature. [ABSTRACT FROM AUTHOR]
- Published
- 2023
- Full Text
- View/download PDF
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.