• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.9
• /
• pp.4665-4683
• /
• 2019

Previous methods build image annotation model by leveraging three basic dependencies: relations between image and label (image/label), between images (image/image) and between labels (label/label). Even though plenty of researches show that multiple dependencies can work jointly to improve annotation performance, different dependencies actually do not "work jointly" in their diagram, whose performance is largely depending on the result predicted by image/label section. To address this problem, we propose the adaptive attention annotation model (AAAM) to associate these dependencies with the prediction path, which is composed of a series of labels (tags) in the order they are detected. In particular, we optimize the prediction path by detecting the relevant labels from the easy-to-detect to the hard-to-detect, which are found using Binary Cross-Entropy (BCE) and Triplet Margin (TM) losses, respectively. Besides, in order to capture the inforamtion of each label, instead of explicitly extracting regional featutres, we propose the self-attention machanism to implicitly enhance the relevant region and restrain those irrelevant. To validate the effective of the model, we conduct experiments on three well-known public datasets, COCO 2014, IAPR TC-12 and NUSWIDE, and achieve better performance than the state-of-the-art methods.

• Journal of The Korean Astronomical Society
• /
• v.53 no.6
• /
• pp.139-147
• /
• 2020

The sunspot area is a critical physical quantity for assessing the solar activity level; forecasts of the sunspot area are of great importance for studies of the solar activity and space weather. We developed an innovative hybrid model prediction method by integrating the complementary ensemble empirical mode decomposition (CEEMD) and extreme learning machine (ELM). The time series is first decomposed into intrinsic mode functions (IMFs) with different frequencies by CEEMD; these IMFs can be divided into three groups, a high-frequency group, a low-frequency group, and a trend group. The ELM forecasting models are established to forecast the three groups separately. The final forecast results are obtained by summing up the forecast values of each group. The proposed hybrid model is applied to the smoothed monthly mean sunspot area archived at NASA's Marshall Space Flight Center (MSFC). We find a mean absolute percentage error (MAPE) and a root mean square error (RMSE) of 1.80% and 9.75, respectively, which indicates that: (1) for the CEEMD-ELM model, the predicted sunspot area is in good agreement with the observed one; (2) the proposed model outperforms previous approaches in terms of prediction accuracy and operational efficiency.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.19 no.2
• /
• pp.117-125
• /
• 2009

In this paper, we analyze the Contact Center's specific-security characteristics, including the threat model and weakness and study effective security measures focussing on protecting customer's personal information. Also, we establish the information security management system to reduce the possibility of information leakage from the internal employee in advance. As a result, we propose the "Security management model for protecting personal information for customer Contact Center" that complies with current ISO/IEC JTC 1 ISMS 27000 series standards.

• Journal of the Korean Geosynthetics Society
• /
• v.21 no.1
• /
• pp.33-48
• /
• 2022

This paper presents illustrative examples of the application of advanced digital image correlation (DIC) technology in the geotechnical laboratory tests, such as shallow footing test, trapdoor test, retaining wall test, and wide width tensile test on geogrid. The theoretical background of the DIC technique is first introduced together with fundamental equations. Relevant reduced-scale model tests were then performed using standard sand while applying the DIC technique to capture the movement of target materials during tests. A number of different approaches were tried to obtain optimized images that allow efficient tracking of material speckles based on the DIC technique. In order to increase the trackability of soil particles, a mix of dyed and regular sand was used during the model tests while specially devised painted speckles were applied to the geogrid. A series of images taken during tests were automatically processed and analyzed using software named VIC-2D that automatically generates displacements and strains. The soil deformation field and associated failure patterns obtained from the DIC technique for each test were found to compare fairly well with the theoretical ones. Also shown is that the DIC technique can also general strains appropriate to the wide width tensile test on geogrid, It is demonstrated in this study that the advanced DIC technique can be effectively used in monitoring the deformation and strain field during a reduced-scale geotechnical model laboratory test.

• The Journal of Fisheries Business Administration
• /
• v.51 no.4
• /
• pp.47-67
• /
• 2020

I explore that South Korea's major import fishery product markets-frozen hairtail, frozen mackerel, frozen pollock and frozen squid-are integrated by testing whether there is favorable evidence of the law of one price (LOP). Unlike previous studies on the LOP for fishery product markets, I assume non-zero import costs and include them in a trade model. To explore whether LOP holds for major import fishery product markets in South Korea with non-zero import costs, I utilize a non-linear time-series model, Smooth Transition Autoregressive (STAR) model with the sample periods from January in 2002 to December in 2019. I find that the behaviors of home-foreign price (i.e., import price) differentials of all four major import fishery products are non-linear depending on whether trade occurs and favorable evidence of LOP for each import market in South Korea. These findings indicate that each of South Korea's major import fishery product markets is integrated. They imply that the supply of each major import fishery product-frozen hairtail, frozen pollock, frozen mackerel and frozen squid, and their prices are stable even if there is an economic shock on each market. When it comes to trade policy implications, the Korean trade policy including tariffs or quotas against their import countries for the four major import fishery products may not have influences on their price in the markets.

• Structural Engineering and Mechanics
• /
• v.83 no.4
• /
• pp.435-449
• /
• 2022

Constitutive modeling that could reasonably predict and effectively evaluate the post-peak structural behavior while eliminating the mesh-dependency in numerical simulation remains to be developed for general engineering applications. Based on the previous work, a simple one-dimensional modeling procedure is proposed to predict and evaluate the post-peak response, as characterized by the evolution of localized strain field, of a steel member to monotonically uniaxial tension. The proposed model extends the classic one-dimensional softening with localization model as introduced by (Schreyer and Chen 1986) to account for the localization length, and bifurcation and rupture points. The new findings of this research are as follows. Two types of strain-softening functions (bilinear and nonlinear) are proposed for comparison. The new failure criterion corresponding to the constitutive modeling is formulated based on the engineering strain inside the localization zone at rupture. Furthermore, a new mathematical expression is developed, based on the strain rate inside and outside the localization zone, to describe the displacement field at which bifurcation occurs. The model solutions are compared with the experimental data on four low-carbon cylindrical steel bars of different lengths. For engineering applications, the model solutions are also compared to the experimental data of a cylindrical steel bar system (three steel bars arranged in series). It is shown that the bilinear and nonlinear softening models can predict the energy dissipation in the post-peak regime with an average difference of only 4%.

• International conference on construction engineering and project management
• /
• 2022.06a
• /
• pp.814-822
• /
• 2022

The deployment of sensors for Structural Health Monitoring requires a complicated network arrangement, ground truthing, and calibration for validating sensor performance periodically. Any conventional sensor on a structural element is also subjected to static and dynamic vertical loadings in conjunction with other environmental factors, such as brightness, noise, temperature, and humidity. A structural model with strain gauges was built and tested to get realistic sensory information. This paper investigates different deep learning architectures and algorithms, including unsupervised, autoencoder, and supervised methods, to benchmark blind drift calibration methods using deep learning. It involves a fully connected neural network (FCNN), a long short-term memory (LSTM), and a gated recurrent unit (GRU) to address the blind drift calibration problem (i.e., performing calibrations of installed sensors when ground truth is not available). The results show that the supervised methods perform much better than unsupervised methods, such as an autoencoder, when ground truths are available. Furthermore, taking advantage of time-series information, the GRU model generates the most precise predictions to remove the drift overall.

• International Journal of Korean Welding Society
• /
• v.1 no.2
• /
• pp.12-17
• /
• 2001

With the advance of the robotic welding process, procedure optimization that selects the welding procedure and predicts bead width that will be deposited is increased. A major concern involving procedure optimization should define a welding procedure that can be shown to be the best with respect to some standard and chosen combination of process parameters, which give an acceptable balance between production rate and the scope of defects for a given situation. This paper presents a new algorithm to establish a mathematical model f3r predicting bead width through a neural network and multiple regression methods, to understand relationships between process parameters and bead width, and to predict process parameters on bead width for GMA welding process. Using a series of robotic arc welding, additional multi-pass butt welds were carried out in order to verify the performance of the neural network estimator and multiple regression methods as well as to select the most suitable model. The results show that not only the proposed models can predict the bead width with reasonable accuracy and guarantee the uniform weld quality, but also a neural network model could be better than the empirical models.

• Journal of the Semiconductor & Display Technology
• /
• v.22 no.2
• /
• pp.6-10
• /
• 2023

High-potential data can be predicted and stored in the cache to prevent cache misses, thus reducing the processor's request and wait times. As a result, the processor can work non-stop, hiding memory latency. By utilizing the temporal/spatial locality of memory access, the prefetcher introduced to improve the performance of these computers predicts the following memory address will be accessed. We propose a prefetcher that applies the GRU model, which is advantageous for handling time series data. Display the currently accessed address in binary and use it as training data to train the Gated Recurrent Unit model based on the difference (delta) between consecutive memory accesses. Finally, using a GRU model with learned memory access patterns, the proposed data prefetcher predicts the memory address to be accessed next. We have compared the model with the multi-layer perceptron, but our prefetcher showed better results than the Multi-Layer Perceptron.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.46 no.4
• /
• pp.281-293
• /
• 2023

This study addresses the optimal design methodology for switching between active electronically scanned array (AESA) radar operating modes to easily select the necessary information to reduce pilots' cognitive load and physical workload in situations where diverse and complex information is continuously provided. This study presents a procedure for defining a hidden Markov chain model (HMM) for modeling operating mode changes based on time series data on the operating modes of the AESA radar used by pilots while performing mission scenarios with inherent uncertainty. Furthermore, based on a transition probability matrix (TPM) of the HMM, this study presents a mathematical programming model for proposing the optimal structural design of AESA radar operating modes considering the manipulation method of a hands on throttle-and-stick (HOTAS). Fighter pilots select and activate the menu key for an AESA radar operation mode by manipulating the HOTAS's rotary and toggle controllers. Therefore, this study presents an optimization problem to propose the optimal structural design of the menu keys so that the pilot can easily change the menu keys to suit the operational environment.