• The Korean Journal of Applied Statistics
• /
• v.34 no.2
• /
• pp.167-175
• /
• 2021

Semi-supervised learning uses both labeled data and unlabeled data. Recently consistency regularization is very popular in semi-supervised learning. Unsupervised data augmentation (UDA) that uses unlabeled data augmentation is also based on the consistency regularization. The Kullback-Leibler divergence is used for the loss of unlabeled data and cross-entropy for the loss of labeled data through UDA learning. UDA uses techniques such as training signal annealing (TSA) and confidence-based masking to promote performance. In this study, we propose to use Jensen-Shannon divergence instead of Kullback-Leibler divergence, reverse-TSA and not to use confidence-based masking for performance improvement. Through experiment, we show that the proposed technique yields better performance than those of UDA.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2023.05a
• /
• pp.204-205
• /
• 2023

This study studied a prediction model for sending collision alarms for small fishing boats based on semi-supervised learning(SSL). The supervised learning (SL) method requires a large number of labeled data, but the labeling process takes a lot of resources and time. This study used service data collected through a data pipeline linked to 'intelligent maritime traffic information service' and data collected from real-sea experiment. The model accuracy was improved as a result of learning not only real-sea experiment data with labeling determined based on actual user satisfaction but also service data without label determined together.

• Journal of Broadcast Engineering
• /
• v.23 no.3
• /
• pp.351-360
• /
• 2018

Human emotion recognition is a research topic that is receiving continuous attention in computer vision and artificial intelligence domains. This paper proposes a method for classifying human emotions through multiple neural networks based on multi-modal signals which consist of image, landmark, and audio in a wild environment. The proposed method has the following features. First, the learning performance of the image-based network is greatly improved by employing both multi-task learning and semi-supervised learning using the spatio-temporal characteristic of videos. Second, a model for converting 1-dimensional (1D) landmark information of face into two-dimensional (2D) images, is newly proposed, and a CNN-LSTM network based on the model is proposed for better emotion recognition. Third, based on an observation that audio signals are often very effective for specific emotions, we propose an audio deep learning mechanism robust to the specific emotions. Finally, so-called emotion adaptive fusion is applied to enable synergy of multiple networks. The proposed network improves emotion classification performance by appropriately integrating existing supervised learning and semi-supervised learning networks. In the fifth attempt on the given test set in the EmotiW2017 challenge, the proposed method achieved a classification accuracy of 57.12%.

• Korean Journal of Remote Sensing
• /
• v.33 no.5_2
• /
• pp.689-698
• /
• 2017

In this paper, a new classification method based on the combination of semi-supervised learning with spatial similarity of adjacent pixels is presented for crop classification in inaccessible areas. Iterative classification based on semi-supervised learning is applied to extract reliable training data from both the initial classification result with a small number of training data, and classification results of adjacent pixels are also considered to extract new training pixels with less uncertainty. To evaluate the applicability of the proposed method, a case study of the classification of field crops was carried out using multi-temporal Landsat-8 OLI acquired in the Daehongdan region, North Korea. From a case study, the misclassification of crops and forests, and isolated pixels in the initial classification result were greatly reduced by applying the proposed semi-supervised learning method. In addition, the combination of classification results of adjacent pixels for the extraction of new training data led to the great reduction of both misclassification results and isolated pixels, compared to the initial classification and traditional semi-supervised learning results. Therefore, it is expected that the proposed method would be effectively applied to classify areas in which it is difficult to collect sufficient training data.

• Journal of Internet Computing and Services
• /
• v.24 no.6
• /
• pp.91-98
• /
• 2023

Recently, according to development of artificial intelligence, a wide range of industry being automatic and optimized. Also we can find out some research of using supervised learning for deteceting defect of railway in domestic rail industry. However, there are structures other than rails on the track, and the fastener is a device that binds the rail to other structures, and periodic inspections are required to prevent safety accidents. In this paper, we present a method of reducing cost for labeling using semi-supervised and transfer model trained on rail fastener data. We use Resnet50 as the backbone network pretrained on ImageNet. At first we randomly take training data from unlabeled data and then labeled that data to train model. After predict unlabeled data by trained model, we adopted a method of adding the data with the highest probability for each class to the training data by a predetermined size. Futhermore, we also conducted some experiments to investigate the influence of the number of initially labeled data. As a result of the experiment, model reaches 92% accuracy which has a performance difference of around 5% compared to supervised learning. This is expected to improve the performance of the classifier by using relatively few labels without additional labeling processes through the proposed method.

• ETRI Journal
• /
• v.41 no.5
• /
• pp.560-573
• /
• 2019

Two supervised learning algorithms, a basic neural network and a long short-term memory recurrent neural network, are applied to traffic including DDoS attacks. The joint effects of preprocessing methods and hyperparameters for machine learning on performance are investigated. Values representing attack characteristics are extracted from datasets and preprocessed by two methods. Binary classification and two optimizers are used. Some hyperparameters are obtained exhaustively for fast and accurate detection, while others are fixed with constants to account for performance and data characteristics. An experiment is performed via TensorFlow on three traffic datasets. Three scenarios are considered to investigate the effects of learning former traffic on sequential traffic analysis and the effects of learning one dataset on application to another dataset, and determine whether the algorithms can be used for recent attack traffic. Experimental results show that the used preprocessing methods, neural network architectures and hyperparameters, and the optimizers are appropriate for DDoS attack detection. The obtained results provide a criterion for the detection accuracy of attacks.

• Journal of Korea Multimedia Society
• /
• v.25 no.10
• /
• pp.1375-1385
• /
• 2022

In this paper, we propose a semi-supervised domain adaptation solution to deal with practical face recognition (FR) scenarios where a single face image for each target identity (to be recognized) is only available in the training phase. Main goal of the proposed method is to reduce the discrepancy between the target and the source domain face images, which ultimately improves FR performances. The proposed method is based on the Domain Adatation network (DAN) using an MMD loss function to reduce the discrepancy between domains. In order to train more effectively, we develop a novel loss function learning strategy in which MMD loss and cross-entropy loss functions are adopted by using different weights according to the progress of each epoch during the learning. The proposed weight adoptation focuses on the training of the source domain in the initial learning phase to learn facial feature information such as eyes, nose, and mouth. After the initial learning is completed, the resulting feature information is used to training a deep network using the target domain images. To evaluate the effectiveness of the proposed method, FR performances were evaluated with pretrained model trained only with CASIA-webface (source images) and fine-tuned model trained only with FERET's gallery (target images) under the same FR scenarios. The experimental results showed that the proposed semi-supervised domain adaptation can be improved by 24.78% compared to the pre-trained model and 28.42% compared to the fine-tuned model. In addition, the proposed method outperformed other state-of-the-arts domain adaptation approaches by 9.41%.

• The Journal of the Acoustical Society of Korea
• /
• v.40 no.6
• /
• pp.578-586
• /
• 2021

We propose a semi-supervised learning method based on Variational AutoEncoder (VAE) and Unsupervised Data Augmentation (UDA) to improve the performance of an end-to-end speech recognizer. In the proposed method, first, the VAE-based augmentation model and the baseline end-to-end speech recognizer are trained using the original speech data. Then, the baseline end-to-end speech recognizer is trained again using data augmented from the learned augmentation model. Finally, the learned augmentation model and end-to-end speech recognizer are re-learned using the UDA-based semi-supervised learning method. As a result of the computer simulation, the augmentation model is shown to improve the Word Error Rate (WER) of the baseline end-to-end speech recognizer, and further improve its performance by combining it with the UDA-based learning method.

• Journal of the Korean Society of Radiology
• /
• v.14 no.7
• /
• pp.991-1001
• /
• 2020

Supervised deep learning technologies for improving the image quality of computed tomography (CT) need a lot of training data. When input images have different characteristics with training images, the technologies cause structural distortion in output images. In this study, an imaging model based on the deep reinforcement learning (DRL) was developed for overcoming the drawbacks of the supervised deep learning technologies and reducing noise in CT images. The DRL model was consisted of shared, value and policy networks, and the networks included convolutional layers, rectified linear unit (ReLU), dilation factors and gate rotation unit (GRU) in order to extract noise features from CT images and improve the performance of the DRL model. Also, the quality of the CT images obtained by using the DRL model was compared to that obtained by using the supervised deep learning model. The results showed that the image accuracy for the DRL model was higher than that for the supervised deep learning model, and the image noise for the DRL model was smaller than that for the supervised deep learning model. Also, the DRL model reduced the noise of the CT images, which had different characteristics with training images. Therefore, the DRL model is able to reduce image noise as well as maintain the structural information of CT images.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.24 no.2
• /
• pp.225-230
• /
• 2020

Machine learning is an algorithm which learns a computer based on the data so that the computer can identify the trend of the data and predict the output of new input data. Machine learning can be classified into supervised learning, unsupervised learning, and reinforcement learning. Supervised learning is a way of learning a machine with given label of data. In other words, a method of inferring a function of the system through a pair of data and a label is used to predict a result using a function inferred about new input data. If the predicted value is continuous, regression analysis is used. If the predicted value is discrete, it is used as a classification. A result of analysis, no. 8 (5, 3.4, setosa), 27 (5, 3.4, setosa), 41 (5, 3.5, setosa), 44 (5, 3.5, setosa) and 40 (5.1, 3.4, setosa) in Table 3 were classified as the most similar Iris flower. Therefore, theoretical practical are suggested.