Uiduk University

Although a number of supervised learning strategies have been reported with promising results, the application of the strategies in denoising digital tomosynthesis (DT) images is limited due to its insufficient performance and a lack of training images. In this study, a multi-agent reinforcement learning (MARL) network was proposed to provide a denoising model for DT images and overcome the drawbacks of the supervised learning. The network consisted of three sub-networks, and, among them, the shared sub-network was variously constructed by using general convolution layers, dilated convolution layers and residual blocks for investigating the effect of feature extraction methods in an output image. The result showed that the proposed model dramatically improved the signal-to-noise ratio (SNR) and structural similarity (SSIM) of output DT images compared to an input image, and the SNRs of the models using the residual blocks were 8.58 to 43.53% higher than those of the other models. The shared sub-networks with the residual blocks improved the spatial frequencies at 10% modulation transfer function (MTF) by 11.26 to 19.49% compared to the other models. In conclusion, the proposed MARL-based denoising model is able to efficiently reduce DT image noise, and the modification of the shared sub-network can improve the performance of the MARL-based DT denoising model.

Purpose:The primary aim of this study is to investigate the potential effects of self-stretching exercises on immediate improvements in cognitive function, with a specific focus on attention and memory.Methods: Thirty healthy subjects participated in the study, meeting the inclusion criteria of having no musculoskeletal or nervous system diseases and the ability to understand and follow the experiment.Cervical range of motion (ROM) was evaluated.Muscle tone and stiffness were measured using the MyotonPRO system.The self-stretching exercises targeted the upper trapezius, levator scapulae, and sternocleidomastoid muscles.Cognitive function was assessed using the Stroop test for selective attention and a memory screening test.The participants underwent baseline assessments followed by education on self-stretching exercises, after which the assessments were repeated.Data analysis included averaging the results from three repeated trials and performing paired t-tests to determine the significance of differences before and after the stretching exercise.Statistical analysis was conducted with p-values less than 0.05 considered statistically significant. Results:The results showed significant improvements in cognitive function, cervical ROM, reduced muscle tone, and stiffness following the self-stretching exercises. Conclusion:This study demonstrates that self-stretching exercises can lead to improved not only cervical ROM, reduced muscle tone and stiffness in specific muscles but also potential enhancements in attention regarding cognitive processing speed.

and use it to propose appropriate educational subjects. Methods: This case study was conducted at U university in Gyeongju-si, Gyeongsangbuk-do. To study curricula based on major competence, related documents were analyzed, expert opinions were obtained, and a consumer survey was conducted. Results: U University assigns major competencies for each department, develops subjects accordingly, maps them, and constructs a competency matrix. The four major competencies of physical therapy were defined as musculoskeletal system, nervous system, cardiorespiratory system and sports injury, and physical management abilities. In order to realize these major competencies, U University added subjects to its curriculum. The results of internal and external environmental analysis should be more actively considered to enable changes in subjects at the departmental level. In addition, the setting of major competencies should be subjected to more detailed, specific reviews to maintain the basic principles of major curricula.

Deep neural networks are vulnerable to maliciously crafted inputs called adversarial examples. Research on unprecedented adversarial attacks is significant since it can help strengthen the reliability of neural networks by alarming potential threats against them. However, since existing adversarial attacks disturb models unconditionally, the resulting adversarial examples increase their detectability through statistical observations or human inspection. To tackle this limitation, we propose hidden conditional adversarial attacks whose resultant adversarial examples disturb models only if the input images satisfy attackers’ pre-defined conditions. These hidden conditional adversarial examples have better stealthiness and controllability of their attack ability. Our experimental results on the CIFAR-10 and ImageNet datasets show their effectiveness and raise a serious concern about the vulnerability of CNNs against the novel attacks.

Deep neural networks have shown outstanding performance in various areas, but adversarial examples can easily fool them. Although strong adversarial attacks have defeated diverse adversarial defense methods, adversarial training, which augments training data with adversarial examples, remains an effective defense strategy. To further improve adversarial robustness, this paper exploits adversarial examples of adversarial examples. We observe that these doubly adversarial examples tend to return to the original prediction on the clean images but sometimes drift toward other classes. From this finding, we propose a regularization loss that prevents these drifts, which mitigates the vulnerability against multi-targeted attacks. Experimental results on the CIFAR-10 and CIFAR-100 datasets empirically show that the proposed loss improves adversarial robustness.

Abstract Kori 1, the first commercial reactor in Korea, was permanently shut down in 2017 and is preparing for decommissioning in the mid‐2020s. Although Kori 1 is shut down, some radiation sources remain in the system; as such, the dismantling workers are exposed to radiation until decommissioning is completed. Unexpected radiation events have been reported in some decommissioning nuclear power plants (NPPs), which resulted in additional radiation exposure to workers and significant financial losses due to delays in the decommissioning schedule. Hence, an appropriate radiation protection program must be established to minimize radiation exposure as well as the economic burden. In this study, radiation doses from precedent decommissioning NPPs worldwide are compared with those estimated from Kori 1 to determine dose reduction measures that can achieve radiation exposure as low as reasonably achievable to decommissioning workers.

Envelope analysis is a widely used tool for fault detection in rotating machines. In envelope analysis, impulsive noise contaminates the measured signal, making it difficult to extract the features of defects. This paper proposes a time–frequency envelope analysis that overcomes the effects of impulsive noises. Envelope analysis is performed by dividing the signal into several sections through a time window. The effect of impulsive noises is eliminated by using the frequency characteristics of the short time rectangular wave. The proposed method was verified through simulation and experimental data. The simulation was conducted by mathematically modeling a cyclo-stationary process that characterizes rotating machinery signals. In addition, the effectiveness of the method was verified by the measured data of normal and defective air-conditioners produced on the actual assembly line. This simple proposed method is effective enough to detect the faults. In the future, the approaches of big data and deep learning will be required for the development of the prognostic health-management framework.

In this study, we present an optimized method for remediating multiple metals and metalloids using pumice permeable reactive barrier with zero-valent iron (ZVI) and with a modified reactor packing bed. The objective was to determine the effect of the contaminant breakthrough capacity of the modified bed, volumetric discharge of treated water, and the reactive material. Arsenic, manganese, iron, and zinc species with the initial concentrations of 0.001, 275.31, 0.61, and 0.063 mmol/L were used as the reference contaminants, respectively. ZVI and pumice were studied to remove the contaminants from synthetic groundwater. Contaminant breakthrough capacity (removal efficiency of 1) was not observed in the ZVI reactor bed with pumice for the 90-day period. The application of irregular reactive bed packing enhanced gas and water transport and removal of heavy metal(loid) for ZVI and pumice column reactor. Although contaminant breakthrough capacity (removal efficiency of 1) was not observed in the ZVI column reactor, clogging phenomena was characteristic. The column with only pumice exhibited zinc and manganese breakthrough capacity after 8 days owing to poor adsorption characteristics because the dominant remediation mechanism was cation exchange. Reactive material characteristics, and the remediation mechanism and for each reactive system, are described in this study. This study showed that the asymmetrical bed geometry could aid in the remediation process and venting of gas buildup.

The device performance deterioration mechanism caused by the total ionizing dose effect after the γ-ray irradiation was investigated in GaN-based metal-insulator-semiconductor high electron mobility transistors (MIS-HEMTs) for a 5 nm-thick SiN and HfO2 gate dielectric layer. The γ-ray radiation hardness according to the gate dielectric layer was also compared between the two different GaN-based MIS-HEMTs. Although HfO2 has exhibited strong tolerance to the total ionizing dose effect in Si-based devices, there is no detail report of the γ-ray radiation effects in GaN-based MIS-HEMTs employing a HfO2 gate dielectric layer. The pulsed-mode stress measurement results and carrier mobility behavior revealed that the device properties not only have direct current (DC) characteristics, but radio frequency (RF) performance has also been mostly degraded by the deterioration of the gate dielectric quality and the trapped charges inside the gate insulator. We also figured out that the immunity to the γ-ray radiation was improved when HfO2 was employed instead of SiN as a gate dielectric layer due to its stronger endurance to the γ-ray irradiation. Our results highlight that the application of a gate insulator that shows superior immunity to the γ-ray irradiation is a crucial factor for the improvement of the total ionizing dose effect in GaN-based MIS-HEMTs.

The construction of artificial weirs as part of the Four Major Rivers Restoration Project is expected to result in the Nakdong River Basin having the characteristics of closed water bodies and properties of water quality that convert suspended components into soluble ones due to a long retention time, thereby leading to the continuous deterioration of water quality. In this regard, this study investigated the changes of water quality on four weirs in the mid-lower Nakdong River Basin. In addition, it examined the organic matters and nutrient salts affecting the water quality, established mass balance with respect to the nutrient salts, and analyzed the load accumulations of each weir to present the directions for water quality management. With regard to the water quality, the items of BOD and COD, T-N and NH4-N, and T-P and PO4-P showed similar tendencies, and most of the items exhibited high concentrations from July to September. Furthermore, it was confirmed that the concentrations of the water quality items were high in Dalseong Weir and Hapcheon-Changnyeong Weir. In July, most of the weirs showed the greatest difference of DO in the surface layer and deepwater layer, while the DO concentration of the deepwater layer dropped below 4.0 mg/L in all four weirs. The sediment survey showed that the SS, T-N, and T-P have a larger amount of sediment in the aphotic zone than in the photic zone, and the amount of sediment per unit area was greatest in Dalseong Weir. According to the sediment release survey results, the release concentration was high in July, which was due to the effects of high water temperature and low DO concentration. T-N was eluted mostly from Dalseong Weir and Gangjeong-Goryeong Weir, and T-P was eluted mostly from Dalseong Weir and Hapcheon-Changnyeong Weir. Among the mass balance, nitrogen balance showed the highest amount of accumulation in Gangjeong-Goryeong Weir (6109 kg/d), and phosphorus balance showed the highest amount of accumulation in Changnyeong-Haman Weir (447 kg/d). These results are expected to be utilized as important materials for setting the management directions for each weir.