• Title/Summary/Keyword: Self-attention mechanism

Search Result 46, Processing Time 0.023 seconds

Generation of High-Resolution Chest X-rays using Multi-scale Conditional Generative Adversarial Network with Attention (주목 메커니즘 기반의 멀티 스케일 조건부 적대적 생성 신경망을 활용한 고해상도 흉부 X선 영상 생성 기법)

  • Ann, Kyeongjin;Jang, Yeonggul;Ha, Seongmin;Jeon, Byunghwan;Hong, Youngtaek;Shim, Hackjoon;Chang, Hyuk-Jae
    • Journal of Broadcast Engineering
    • /
    • v.25 no.1
    • /
    • pp.1-12
    • /
    • 2020
  • In the medical field, numerical imbalance of data due to differences in disease prevalence is a common problem. It reduces the performance of a artificial intelligence network, leading to difficulties in learning a network with good performance. Recently, generative adversarial network (GAN) technology has been introduced as a way to address this problem, and its ability has been demonstrated by successful applications in various fields. However, it is still difficult to achieve good results in solving problems with performance degraded by numerical imbalances because the image resolution of the previous studies is not yet good enough and the structure in the image is modeled locally. In this paper, we propose a multi-scale conditional generative adversarial network based on attention mechanism, which can produce high resolution images to solve the numerical imbalance problem of chest X-ray image data. The network was able to produce images for various diseases by controlling condition variables with only one network. It's efficient and effective in that the network don't need to be learned independently for all disease classes and solves the problem of long distance dependency in image generation with self-attention mechanism.

The Effect of Future Entrepreneurs' Marketing Self-efficacy on Entrepreneurial Intention: The Mediating Role of Resilience (예비 창업자의 마케팅 효능감이 창업의지에 미치는 영향: 회복탄력성의 매개역할)

  • Kim, Jung-Rae
    • Journal of Convergence for Information Technology
    • /
    • v.10 no.11
    • /
    • pp.131-140
    • /
    • 2020
  • Entrepreneurial self-efficacy and self-efficacy on the positive effects of new firm creation are well developed in past research. However, marketing self-efficacy has not received enough attention on the prediction of entrepreneurial intention despite marketing and entrepreneurship are crucial when creating a new firm. Moreover, resilience has a central role in entrepreneurship research while present study aims to explore the impact of psychological factors on new venture creation but scholars have not begun to uncover specific mechanism through which marketing self-efficacy, resilience and entrepreneurial intention. The purpose of this study then was to examine resilience as mediator in the marketing self-efficacy and entrepreneurial intention relationship. Questionnaires were employed to collect data from major universities of future entrepreneurs in Korea. A total of 315 completed questionnaires were returned. Results showed that marketing self-efficacy and resilience had a positive effect on entrepreneurial intention. Further, resilience had a significant mediating effect on the marketing self-efficacy and entrepreneurial intention relationship. These results suggest theoretical and practical implications as an important factor to stimulate entrepreneurial intention.

Effects of online game addiction on impulsive crime through Self-control (온라인 게임 중독 특성이 자아통제를 통해 충동성 범죄에 미치는 영향)

  • Lee, Jiyeop;Kwon, Dosoon
    • Journal of Korea Society of Digital Industry and Information Management
    • /
    • v.13 no.1
    • /
    • pp.135-145
    • /
    • 2017
  • Distribution of online media sharply accelerated in the modern society, leading to robust increase in the time where individuals spend their leisure time on the Internet. One of the most popular entertainments among this is the video game industry, which gradually came to possess the characteristics like violence, cruelty, gambling, and sexuality, in order to gain attention. This study aims to investigate the addiction characteristics of users of domestic online games and to verify the causal relationship of these factors on the factors that affect impulsive crimes through Self-control. For practical verification of this study, survey was conducted for students of W High School in Incheon. This study is meaningful in that it prepared a theoretical foundation in the process of inferring criminal intent for impulsive crimes based on the effects of online games on impulsiveness, aggression, and addiction. It is also significant in that it presented a new mechanism that ego control, which occurred in the process of presenting a new variable of negative effect of online game on the criminal intent of impulsive crime, affects impulsive crime.

The Influence of Self-Construal on Conditionalization and Discounting Effect in Contingency Judgment (수반성 판단에서 자기해석이 조건부화와 절감효과에 미치는 영향)

  • Kim, Kyungil;Kim, Tae Hoon
    • Korean Journal of Cognitive Science
    • /
    • v.24 no.4
    • /
    • pp.323-338
    • /
    • 2013
  • There are multiple process mechanism in causal reasoning, which is estimation of the causal strength between cause and result. Further, because these mechanisms operate on different time phase during causal reasoning, it is highly possible that different individual difference factors are related to individual mechanisms of causal reasoning. Especially, the phenomena of conditionalization and discounting reflect attention to multiple potential causes when people infer the relationship between cause and effect. In this study, we manipulated self-construal which is an individual difference factor that reflects context sensitivity in cognition. As results, no difference was observed in conditionalization between individuals with an independent self-construal and those with an interdependent self-construal. However, independent self-construal group was observed to be lower in discounting than the interdependent self-construal group. The results indicate that conditionalization and discounting are independent cognitive process in human causal reasoning.

  • PDF

The effect of flue-gas recirculation on combustion characteristics of regenerative low NOx burner (축열식 저 NOx 연소기의 배가스 재순환이 연소특성에 미치는 영향)

  • Kang, Min-Wook;Yoon, Young-Bin;Dong, Sang-Keun
    • 한국연소학회:학술대회논문집
    • /
    • 2002.11a
    • /
    • pp.97-104
    • /
    • 2002
  • The conventional regenerative system has a high thermal efficiency as well as energy saving using the high preheated combustion air. in spite of these advantages, it can not avoid high nitric oxide emissions. Recently, flameless combustion has received much attention to solve these problems. In this research, numerical analysis is performed for flow-combustion phenomena in the self regenerative burner. In this analysis we used Fluent 6.0 code. the that is developed for commercial use, Methane gas is used as a fuel and two-step reaction model for methane and Zeldovich mechanism for NO generation are used. the velocity of the preheated combustion air is used as a parameter and we analyze the characteristics of flow-field, temperature distributions and NO emissions. Due to the increased recirculation rate, the maximum temperature of flame is significantly increased and NOx emissions is reduced

  • PDF

Incorporating BERT-based NLP and Transformer for An Ensemble Model and its Application to Personal Credit Prediction

  • Sophot Ky;Ju-Hong Lee;Kwangtek Na
    • Smart Media Journal
    • /
    • v.13 no.4
    • /
    • pp.9-15
    • /
    • 2024
  • Tree-based algorithms have been the dominant methods used build a prediction model for tabular data. This also includes personal credit data. However, they are limited to compatibility with categorical and numerical data only, and also do not capture information of the relationship between other features. In this work, we proposed an ensemble model using the Transformer architecture that includes text features and harness the self-attention mechanism to tackle the feature relationships limitation. We describe a text formatter module, that converts the original tabular data into sentence data that is fed into FinBERT along with other text features. Furthermore, we employed FT-Transformer that train with the original tabular data. We evaluate this multi-modal approach with two popular tree-based algorithms known as, Random Forest and Extreme Gradient Boosting, XGBoost and TabTransformer. Our proposed method shows superior Default Recall, F1 score and AUC results across two public data sets. Our results are significant for financial institutions to reduce the risk of financial loss regarding defaulters.

Exploration of growth mechanism for layer controllable graphene on copper

  • Song, Woo-Seok;Kim, Yoo-Seok;Kim, Soo-Youn;Kim, Sung-Hwan;Jung, Dae-Sung;Jun, Woo-Sung;Jeon, Cheol-Ho;Park, Chong-Yun
    • Proceedings of the Korean Vacuum Society Conference
    • /
    • 2011.02a
    • /
    • pp.490-490
    • /
    • 2011
  • Graphene, hexagonal network of carbon atoms forming a one-atom thick planar sheet, has been emerged as a fascinating material for future nanoelectronics. Huge attention has been captured by its extraordinary electronic properties, such as bipolar conductance, half integer quantum Hall effect at room temperature, ballistic transport over ${\sim}0.4{\mu}m$ length and extremely high carrier mobility at room temperature. Several approaches have been developed to produce graphene, such as micromechanical cleavage of highly ordered pyrolytic graphite using adhesive tape, chemical reduction of exfoliated graphite oxide, epitaxial growth of graphene on SiC and single crystalline metal substrate, and chemical vapor deposition (CVD) synthesis. In particular, direct synthesis of graphene using metal catalytic substrate in CVD process provides a new way to large-scale production of graphene film for realization of graphene-based electronics. In this method, metal catalytic substrates including Ni and Cu have been used for CVD synthesis of graphene. There are two proposed mechanism of graphene synthesis: carbon diffusion and precipitation for graphene synthesized on Ni, and surface adsorption for graphene synthesized on Cu, namely, self-limiting growth mechanism, which can be divided by difference of carbon solubility of the metals. Here we present that large area, uniform, and layer controllable graphene synthesized on Cu catalytic substrate is achieved by acetylene-assisted CVD. The number of graphene layer can be simply controlled by adjusting acetylene injection time, verified by Raman spectroscopy. Structural features and full details of mechanism for the growth of layer controllable graphene on Cu were systematically explored by transmission electron microscopy, atomic force microscopy, and secondary ion mass spectroscopy.

  • PDF

DATCN: Deep Attention fused Temporal Convolution Network for the prediction of monitoring indicators in the tunnel

  • Bowen, Du;Zhixin, Zhang;Junchen, Ye;Xuyan, Tan;Wentao, Li;Weizhong, Chen
    • Smart Structures and Systems
    • /
    • v.30 no.6
    • /
    • pp.601-612
    • /
    • 2022
  • The prediction of structural mechanical behaviors is vital important to early perceive the abnormal conditions and avoid the occurrence of disasters. Especially for underground engineering, complex geological conditions make the structure more prone to disasters. Aiming at solving the problems existing in previous studies, such as incomplete consideration factors and can only predict the continuous performance, the deep attention fused temporal convolution network (DATCN) is proposed in this paper to predict the spatial mechanical behaviors of structure, which integrates both the temporal effect and spatial effect and realize the cross-time prediction. The temporal convolution network (TCN) and self-attention mechanism are employed to learn the temporal correlation of each monitoring point and the spatial correlation among different points, respectively. Then, the predicted result obtained from DATCN is compared with that obtained from some classical baselines, including SVR, LR, MLP, and RNNs. Also, the parameters involved in DATCN are discussed to optimize the prediction ability. The prediction result demonstrates that the proposed DATCN model outperforms the state-of-the-art baselines. The prediction accuracy of DATCN model after 24 hours reaches 90 percent. Also, the performance in last 14 hours plays a domain role to predict the short-term behaviors of the structure. As a study case, the proposed model is applied in an underwater shield tunnel to predict the stress variation of concrete segments in space.

Pedestrian and Vehicle Distance Estimation Based on Hard Parameter Sharing (하드 파라미터 쉐어링 기반의 보행자 및 운송 수단 거리 추정)

  • Seo, Ji-Won;Cha, Eui-Young
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.26 no.3
    • /
    • pp.389-395
    • /
    • 2022
  • Because of improvement of deep learning techniques, deep learning using computer vision such as classification, detection and segmentation has also been used widely at many fields. Expecially, automatic driving is one of the major fields that applies computer vision systems. Also there are a lot of works and researches to combine multiple tasks in a single network. In this study, we propose the network that predicts the individual depth of pedestrians and vehicles. Proposed model is constructed based on YOLOv3 for object detection and Monodepth for depth estimation, and it process object detection and depth estimation consequently using encoder and decoder based on hard parameter sharing. We also used attention module to improve the accuracy of both object detection and depth estimation. Depth is predicted with monocular image, and is trained using self-supervised training method.

Detecting Complex 3D Human Motions with Body Model Low-Rank Representation for Real-Time Smart Activity Monitoring System

  • Jalal, Ahmad;Kamal, Shaharyar;Kim, Dong-Seong
    • KSII Transactions on Internet and Information Systems (TIIS)
    • /
    • v.12 no.3
    • /
    • pp.1189-1204
    • /
    • 2018
  • Detecting and capturing 3D human structures from the intensity-based image sequences is an inherently arguable problem, which attracted attention of several researchers especially in real-time activity recognition (Real-AR). These Real-AR systems have been significantly enhanced by using depth intensity sensors that gives maximum information, in spite of the fact that conventional Real-AR systems are using RGB video sensors. This study proposed a depth-based routine-logging Real-AR system to identify the daily human activity routines and to make these surroundings an intelligent living space. Our real-time routine-logging Real-AR system is categorized into two categories. The data collection with the use of a depth camera, feature extraction based on joint information and training/recognition of each activity. In-addition, the recognition mechanism locates, and pinpoints the learned activities and induces routine-logs. The evaluation applied on the depth datasets (self-annotated and MSRAction3D datasets) demonstrated that proposed system can achieve better recognition rates and robust as compare to state-of-the-art methods. Our Real-AR should be feasibly accessible and permanently used in behavior monitoring applications, humanoid-robot systems and e-medical therapy systems.