• Proceedings of the Korea Database Society Conference
• /
• 1999.06a
• /
• pp.175-186
• /
• 1999

Detecting the features of significant patterns from their own historical data is so much crucial to good performance specially in time-series forecasting. Recently, a new data filtering method (or multi-scale decomposition) such as wavelet analysis is considered more useful for handling the time-series that contain strong quasi-cyclical components than other methods. The reason is that wavelet analysis theoretically makes much better local information according to different time intervals from the filtered data. Wavelets can process information effectively at different scales. This implies inherent support fer multiresolution analysis, which correlates with time series that exhibit self-similar behavior across different time scales. The specific local properties of wavelets can for example be particularly useful to describe signals with sharp spiky, discontinuous or fractal structure in financial markets based on chaos theory and also allows the removal of noise-dependent high frequencies, while conserving the signal bearing high frequency terms of the signal. To date, the existing studies related to wavelet analysis are increasingly being applied to many different fields. In this study, we focus on several wavelet thresholding criteria or techniques to support multi-signal decomposition methods for financial time series forecasting and apply to forecast Korean Won / U.S. Dollar currency market as a case study. One of the most important problems that has to be solved with the application of the filtering is the correct choice of the filter types and the filter parameters. If the threshold is too small or too large then the wavelet shrinkage estimator will tend to overfit or underfit the data. It is often selected arbitrarily or by adopting a certain theoretical or statistical criteria. Recently, new and versatile techniques have been introduced related to that problem. Our study is to analyze thresholding or filtering methods based on wavelet analysis that use multi-signal decomposition algorithms within the neural network architectures specially in complex financial markets. Secondly, through the comparison with different filtering techniques' results we introduce the present different filtering criteria of wavelet analysis to support the neural network learning optimization and analyze the critical issues related to the optimal filter design problems in wavelet analysis. That is, those issues include finding the optimal filter parameter to extract significant input features for the forecasting model. Finally, from existing theory or experimental viewpoint concerning the criteria of wavelets thresholding parameters we propose the design of the optimal wavelet for representing a given signal useful in forecasting models, specially a well known neural network models.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.1
• /
• pp.140-145
• /
• 2010

In this work, to theoretically analyze the nonlinear charging characteristic, a Radial Basis Function Neural Network (RBFNN) is adopted. Based on the RBFNN, an charging characteristic tendency of a Linear Type Magnetic Flux Pump (LTMFP) is analyzed. In the paper, we developed the LTMFP that generates stable and controllable charging current and also experimentally investigated its charging characteristic in the cryogenic system. From these experimental results, the charging current of the LTMFP was also found to be frequency dependent with nonlinear quality due to the nonlinear magnetic behaviour of superconducting Nb foil. On the whole, in the case of essentially cryogenic experiment, since cooling costs loomed large in the cryogenic environment, it is difficult to carry out various experiments. Consequentially, in this paper, we estimated the nonlinear characteristic of charging current as well as realized the intelligent model via the design of RBFNN based on the experimental data. In this paper, we view RBF neural networks as predominantly data driven constructs whose processing is based upon an effective usage of experimental data through a prudent process of Fuzzy C-Means clustering method. Also, the receptive fields of the proposed RBF neural network are formed by the FCM clustering.

• Journal of Broadcast Engineering
• /
• v.26 no.2
• /
• pp.125-131
• /
• 2021

In recent years, Convolutional Neural Networks (CNNs) have achieved outstanding performance in the fields of computer vision such as image classification, object detection, visual quality enhancement, etc. However, as huge amount of computation and memory are required in CNN models, there is a limitation in the application of CNN to low-power environments such as mobile or IoT devices. Therefore, the need for neural network compression to reduce the model size while keeping the task performance as much as possible has been emerging. In this paper, we propose a method to compress CNN models by combining matrix decomposition methods of LR (Low-Rank) approximation and CP (Canonical Polyadic) decomposition. Unlike conventional methods that apply one matrix decomposition method to CNN models, we selectively apply two decomposition methods depending on the layer types of CNN to enhance the compression performance. To evaluate the performance of the proposed method, we use the models for image classification such as VGG-16, RestNet50 and MobileNetV2 models. The experimental results show that the proposed method gives improved classification performance at the same range of 1.5 to 12.1 times compression ratio than the existing method that applies only the LR approximation.

• The Journal of the Korea Contents Association
• /
• v.21 no.3
• /
• pp.616-625
• /
• 2021

Deep neural networks are an approximation method that approximates an arbitrary function to a linear model and then repeats additional approximation using a nonlinear active function. In this process, the method of evaluating the performance of approximation uses the loss function. Existing in-depth learning methods implement approximation that takes into account loss functions in the linear approximation process, but non-linear approximation phases that use active functions use non-linear transformation that is not related to reduction of loss functions of loss. This study proposes parametric activation functions that introduce scale parameters that can change the scale of activation functions and location parameters that can change the location of activation functions. By introducing parametric activation functions based on scale and location parameters, the performance of nonlinear approximation using activation functions can be improved. The scale and location parameters in each hidden layer can improve the performance of the deep neural network by determining parameters that minimize the loss function value through the learning process using the primary differential coefficient of the loss function for the parameters in the backpropagation. Through MNIST classification problems and XOR problems, parametric activation functions have been found to have superior performance over existing activation functions.

• Composites Research
• /
• v.36 no.1
• /
• pp.48-52
• /
• 2023

Fiber-reinforced composites have anisotropic material properties, so the mechanical properties of composite structures can vary depending on the stacking sequence. Therefore, it is essential to design the proper stacking sequence of composite structures according to the functional requirements. However, depending on the manufacturing condition or the shape of the structure, there are many cases where the designed stacking angle is out of range, which can affect structural performance. Accordingly, it is important to analyze the stacking angle in order to confirm that the composite structure is correctly fabricated as designed. In this study, the stacking angle was predicted from real cross-sectional images of fiber-reinforced composites using convolutional neural network (CNN)-based deep learning. Carbon fiber-reinforced composite specimens with several stacking angles were fabricated and their cross-sections were photographed on a micro-scale using an optical microscope. The training was performed for a CNN-based deep learning model using the cross-sectional image data of the composite specimens. As a result, the stacking angle can be predicted from the actual cross-sectional image of the fiber-reinforced composite with high accuracy.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.2
• /
• pp.193-200
• /
• 2023

Cryptanalysis can be performed by various techniques such as known plaintext attack, differential attack, side-channel analysis, and the like. Recently, many studies have been conducted on cryptanalysis using deep learning. A known-plaintext attack is a technique that uses a known plaintext and ciphertext pair to find a key. In this paper, we use deep learning technology to perform a known-plaintext attack against S-PRESENT, a reduced version of the lightweight block cipher PRESENT. This paper is significant in that it is the first known-plaintext attack based on deep learning performed on a reduced lightweight block cipher. For cryptanalysis, MLP (Multi-Layer Perceptron) and 1D and 2D CNN(Convolutional Neural Network) models are used and optimized, and the performance of the three models is compared. It showed the highest performance in 2D convolutional neural networks, but it was possible to attack only up to some key spaces. From this, it can be seen that the known-plaintext attack through the MLP model and the convolutional neural network is limited in attackable key bits.

• Korean Journal of Remote Sensing
• /
• v.39 no.2
• /
• pp.143-156
• /
• 2023

To precisely and reliably analyze the contents of the satellite imagery, recognizing the clouds which are the obstacle to gathering the useful information is essential. In recent times, deep learning yielded satisfactory results in various tasks, so many studies using deep neural networks have been conducted to improve the performance of cloud detection. However, existing methods for cloud detection have the limitation on increasing the performance due to the adopting the network models for semantic image segmentation without modification. To tackle this problem, we introduced the multi-branch neural architecture search to find optimal network structure for cloud detection. Additionally, the proposed method adopts the soft intersection over union (IoU) as loss function to mitigate the disagreement between the loss function and the evaluation metric and uses the various data augmentation methods. The experiments are conducted using the cloud detection dataset acquired by Arirang-3/3A satellite imagery. The experimental results showed that the proposed network which are searched network architecture using cloud dataset is 4% higher than the existing network model which are searched network structure using urban street scenes with regard to the IoU. Also, the experimental results showed that the soft IoU exhibits the best performance on cloud detection among the various loss functions. When comparing the proposed method with the state-of-the-art (SOTA) models in the field of semantic segmentation, the proposed method showed better performance than the SOTA models with regard to the mean IoU and overall accuracy.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1032-1032
• /
• 2001

Artificial Neural Network (ANN) calibration techniques have been used commercially for agricultural applications since the mid-nineties. Global models, based on transmission data from 850 to 1050 nm, are used routinely to measure protein and moisture in wheat and barley and also moisture in triticale, rye, and oats. These models are currently used commercially in approx. 15 countries throughout the world. Results concerning earlier European ANN models are being published elsewhere. Some of the findings from that study will be discussed here. ANN models have also been developed for coarsely ground samples of compound feed and feed ingredients, again measured in transmission mode from 850 to 1050 nm. The performance of models for pig- and poultry feed will be discussed briefly. These models were developed from a very large data set (more than 20,000 records), and cover a very broad range of finished products. The prediction curves are linear over the entire range for protein, fat moisture, fibre, and starch (measured only on poultry feed), and accuracy is in line with the performance of smaller models based on Partial Least Squares (PLS). A simple bias adjustment is sufficient for calibration transfer across instruments. Recently, we have investigated the possible use of ANN for a different type of NIR spectrometer, based on reflectance data from 1100 to 2500 nm. In one study, based on data for protein, fat, and moisture measured on unground compound feed samples, dedicated ANN models for specific product classes (cattle feed, pig feed, broiler feed, and layers feed) gave moderately better Standard Errors of Prediction (SEP) compared to modified PLS (MPLS). However, if the four product classes were combined into one general calibration model, the performance of the ANN model deteriorated only slightly compared to the class-specific models, while the SEP values for the MPLS predictions doubled. Brix value in molasses is a measure of sugar content. Even with a huge dataset, PLS models were not sufficiently accurate for commercial use. In contrast an ANN model based on the same data improved the accuracy considerably and straightened out non-linearity in the prediction plot. The work of Mr. David Funk (GIPSA, U. S. Department of Agriculture) who has studied the influence of various types of spectral distortions on ANN- and PLS models, thereby providing comparative information on the robustness of these models towards instrument differences, will be discussed. This study was based on data from different classes of North American wheat measured in transmission from 850 to 1050 nm. The distortions studied included the effect of absorbance offset pathlength variation, presence of stray light bandwidth, and wavelength stretch and offset (either individually or combined). It was shown that a global ANN model was much less sensitive to most perturbations than class-specific GIPSA PLS calibrations. It is concluded that ANN models based on large data sets offer substantial advantages over PLS models with respect to accuracy, range of materials that can be handled by a single calibration, stability, transferability, and sensitivity to perturbations.

• Journal of Digital Convergence
• /
• v.17 no.1
• /
• pp.239-247
• /
• 2019

An accurate prediction of emotion is a very important issue for the sake of patient-centered medical device development and emotion-related psychology fields. Although there have been many studies on emotion prediction, no studies have applied the heart rate variability and neuro-fuzzy approach to emotion prediction. We propose ANFEP(Adaptive Neuro Fuzzy System for Emotion Prediction) HRV. The ANFEP bases its core functions on an ANFIS(Adaptive Neuro-Fuzzy Inference System) which integrates neural networks with fuzzy systems as a vehicle for training predictive models. To prove the proposed model, 50 participants were invited to join the experiment and Heart rate variability was obtained and used to input the ANFEP model. The ANFEP model with STDRR and RMSSD as inputs and two membership functions per input variable showed the best results. The result out of applying the ANFEP to the HRV metrics proved to be significantly robust when compared with benchmarking methods like linear regression, support vector regression, neural network, and random forest. The results show that reliable prediction of emotion is possible with less input and it is necessary to develop a more accurate and reliable emotion recognition system.

• KIPS Transactions on Software and Data Engineering
• /
• v.12 no.4
• /
• pp.179-188
• /
• 2023

Recently, fake news disguises the form of news content and appears whenever important events occur, causing social confusion. Accordingly, artificial intelligence technology is used as a research to detect fake news. Fake news detection approaches such as automatically recognizing and blocking fake news through natural language processing or detecting social media influencer accounts that spread false information by combining with network causal inference could be implemented through deep learning. However, fake news detection is classified as a difficult problem to solve among many natural language processing fields. Due to the variety of forms and expressions of fake news, the difficulty of feature extraction is high, and there are various limitations, such as that one feature may have different meanings depending on the category to which the news belongs. In this paper, emotional change patterns are presented as an additional identification criterion for detecting fake news. We propose a model with improved performance by applying a convolutional neural network to a fake news data set to perform analysis based on content characteristics and additionally analyze emotional change patterns. Sentimental polarity is calculated for the sentences constituting the news and the result value dependent on the sentence order can be obtained by applying long-term and short-term memory. This is defined as a pattern of emotional change and combined with the content characteristics of news to be used as an independent variable in the proposed model for fake news detection. We train the proposed model and comparison model by deep learning and conduct an experiment using a fake news data set to confirm that emotion change patterns can improve fake news detection performance.