• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.6
• /
• pp.2115-2127
• /
• 2021

Although significant progress has been made in synthesizing visually realistic face images by Generative Adversarial Networks (GANs), there still lacks effective approaches to provide fine-grained control over the generation process for semantic facial attribute editing. In this work, we propose a novel cross channel self-attention based generative adversarial network (CCA-GAN), which weights the importance of multiple channels of features and archives pixel-level feature alignment and conversion, to reduce the impact on irrelevant attributes while editing the target attributes. Evaluation results show that CCA-GAN outperforms state-of-the-art models on the CelebA dataset, reducing Fréchet Inception Distance (FID) and Kernel Inception Distance (KID) by 15~28% and 25~100%, respectively. Furthermore, visualization of generated samples confirms the effect of disentanglement of the proposed model.

• ETRI Journal
• /
• v.44 no.3
• /
• pp.438-449
• /
• 2022

This study proposes a dependency grammar-based self-attention multilayered bidirectional long short-term memory (DG-M-Bi-LSTM) model for subject-predicate-object (SPO) tuple recognition from natural language (NL) sentences. To add recent knowledge to the knowledge base autonomously, it is essential to extract knowledge from numerous NL data. Therefore, this study proposes a high-accuracy SPO tuple recognition model that requires a small amount of learning data to extract knowledge from NL sentences. The accuracy of SPO tuple recognition using DG-M-Bi-LSTM is compared with that using NL-based self-attention multilayered bidirectional LSTM, DG-based bidirectional encoder representations from transformers (BERT), and NL-based BERT to evaluate its effectiveness. The DG-M-Bi-LSTM model achieves the best results in terms of recognition accuracy for extracting SPO tuples from NL sentences even if it has fewer deep neural network (DNN) parameters than BERT. In particular, its accuracy is better than that of BERT when the learning data are limited. Additionally, its pretrained DNN parameters can be applied to other domains because it learns the structural relations in NL sentences.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2021.05a
• /
• pp.343-346
• /
• 2021

약물 디자인이란 단백질과 같은 생물학적 표적에 작용할 수 있는 새로운 약물을 개발하는 과정이다. 전통적인 방법은 탐색과 개발 단계로 구성되어 있으나, 하나의 신약 개발을 위해서는 10 년 이상의 장시간이 요구되기 때문에, 이러한 기간을 단축하기 위한 인공지능 기반의 약물 디자인 방법들이 개발되고 있다. 하지만 많은 심층학습 기반의 약물 디자인 모델들은 RNN 기법을 활용하고 있고, RNN 은 훈련속도가 느리다는 단점이 있기 때문에 개선의 여지가 남아있다. 이런 단점을 극복하기 위해 본 연구는 self-attention 과 variational autoencoder 를 활용한 SMILES 생성 모델을 제안한다. 제안된 모델은 최신 약물 디자인 모델 대비 훈련 시간을 1/36 단축하고, 뿐만 아니라 유효한 SMILES 를 더 많이 생성하는 것을 확인하였다.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.16 no.1
• /
• pp.429-437
• /
• 2015

The purpose of this study was to investigate physical activity level, sleep quality, attention control, and self-regulated learning along to smartphone addiction level among college students. The data were obtained from 269 college students by structured questionnaire, analyzed using SPSS 18.0. The results showed that significant differences with smartphone addictoin level and, gender, grade level, daily using time, physical activity level, sleep quality and attention control. Smartphone addictoin level have correlations with physical activity level, sleep quality, attention control, and self-regulated learning. To prevent the falling off of physical strength and poor school performance, it is needed a strategies for control of smartphone addiction.

• The Journal of the Acoustical Society of Korea
• /
• v.39 no.5
• /
• pp.483-489
• /
• 2020

This paper proposes acoustic model training using self-attention for low-resource speech recognition. In low-resource speech recognition, it is difficult for acoustic model to distinguish certain phones. For example, plosive /d/ and /t/, plosive /g/ and /k/ and affricate /z/ and /ch/. In acoustic model training, the self-attention generates attention weights from the deep neural network model. In this study, these weights handle the similar pronunciation error for low-resource speech recognition. When the proposed method was applied to Time Delay Neural Network-Output gate Projected Gated Recurrent Unit (TNDD-OPGRU)-based acoustic model, the proposed model showed a 5.98 % word error rate. It shows absolute improvement of 0.74 % compared with TDNN-OPGRU model.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.1-16
• /
• 2022

Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928 × 3264 are collected, 160 for training and the rest 40 for the test. 512 × 512 patches are generated from the original images by a sliding window with an overlap of 256 as inputs. Results show that the average IoU between the recognized and ground-truth cracks reaches 0.409, which is 29.8% higher than the regular U-net. A five-fold cross-validation study is performed to verify that the proposed method is robust to different training and test images. Ablation experiments further demonstrate the effectiveness of the proposed SASA neuron and CRED algorithm. Promotions of the average IoU individually utilizing the SASA and CRED module add up to the final promotion of the full model, indicating that the SASA and CRED modules contribute to the different stages of model and data in the training process.

• Journal of Korean Home Economics Education Association
• /
• v.28 no.2
• /
• pp.95-111
• /
• 2016

This study investigated the descriptive statistics and correlation among self-identity, school violence experience, and family social capital of adolescents and examined influence of school violence experience on self-identity and moderating effect of family social capital on the relationship between school violence experience and self-identity. Data used for analysis was from 7th grade students in The Korean Children and Youth Panel Survey in 2012. Analyses were performed using the IBM SPSS program for demographic analysis, pearson correlation, and stepwise regression analyses. Results of the study were as follows: First, the average was slightly higher for self-identity, parents' affectionate attention, and awareness of their child's friends; the average was lower for misconduct experience and victimization experience; second, there was a weak negative correlation between self-identity and bully victimization; there was a positive correlation between self-identity and family social capital (parents' affectionate attention and awareness of their child's friends). Third, to investigate the effect of school violence experience (bullying and bully victimization) on self-identity, stepwise regression analysis results were as follows: Bullying had a statistically positive influence on self-identity and bully victimization had a statistically negative influence on self-identity; both parents' affectionate attention and awareness of their child's friends had a statistically positive influence on self-identity; fourth, parents' affectionate attention had a statistically negative moderating effect on the self-identity; therefore, it signifies that the relationship between bully victimization and self-identity appears differently depending on the parents' affectionate attention, which means that the parents' affectionate attention had a negative effect on the self-identity of the adolescents who were victimized by school violence.

• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.453-461
• /
• 2021

Many deep learning approaches have been studied for image classification in computer vision. However, there are not enough data to generate accurate models in medical fields, and many datasets are not annotated. This study presents a new method that can use both unlabeled and labeled data. The proposed method is applied to classify cervix images into normal versus cancerous, and we demonstrate the results. First, we use a patch self-supervised learning for training the global context of the image using an unlabeled image dataset. Second, we generate a classifier model by using the transferred knowledge from self-supervised learning. We also apply attention learning to capture the local features of the image. The combined method provides better performance than state-of-the-art approaches in accuracy and sensitivity.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.7 no.11
• /
• pp.2824-2838
• /
• 2013

There has been a lot of attention paid recently to analyze dynamic human gestures that vary over time. Most attention to dynamic gestures concerns with spatio-temporal features, as compared to analyzing each frame of gestures separately. For accurate dynamic gesture recognition, motion feature extraction algorithms need to find representative features that uniquely identify time-varying gestures. This paper proposes a new feature-extraction algorithm using temporal self-similarity based on a hierarchical human model. Because a conventional temporal self-similarity method computes a whole movement among the continuous frames, the conventional temporal self-similarity method cannot recognize different gestures with the same amount of movement. The proposed model-based temporal self-similarity method groups body parts of a hierarchical model into several sets and calculates movements for each set. While recognition results can depend on how the sets are made, the best way to find optimal sets is to separate frequently used body parts from less-used body parts. Then, we apply a multiclass support vector machine whose optimization algorithm is based on structural support vector machines. In this paper, the effectiveness of the proposed feature extraction algorithm is demonstrated in an application for taebo gesture recognition. We show that the model-based temporal self-similarity method can overcome the shortcomings of the conventional temporal self-similarity method and the recognition results of the model-based method are superior to that of the conventional method.

• Journal of the Korean Society of School Health
• /
• v.23 no.2
• /
• pp.237-245
• /
• 2010

Purpose: The purpose of this thesis was to examine the effects of social skills training program onto the children with tendency of attention-deficit hyperactivity disorder. Methods: This study used nonequivalent control group pre/post-test quasi-experimental research design. The subjects were 18 children with tendency of attention- deficit hyperactivity in D City. The subjects were divided into two groups, an experimental group of 8 children and a control group of 10. The program consisted of 20 sessions of 60 minutes per session, 5 days a weeks, for 4 weeks. The research tools included Conner's Teacher Rating Scales (CTRS) and Social Skills Rating System (SSRS). The collected data were analyzed using $x^2$ test, Mann-Whitney test on the SPSS 17.0 program. Results: a) the scores for cooperation, self-assertiveness, self-control and empathy increased significantly in the experimental group, compared to the control group. b) the scores for social skills increased significantly in the experimental group, compared to the control group. Conclusion: It appears that the social skills training program is a useful nursing intervention to improve the social skills for children with tendency of attention-deficit hyperactivity.