• International Journal of Computer Science & Network Security
• /
• v.23 no.11
• /
• pp.133-141
• /
• 2023

This work take deeper analysis of Adaptive Moment Estimation (Adam) and Adam with Weight Decay (AdamW) implementation in real world text classification problem (DGA Malware Detection). AdamW is introduced by decoupling weight decay from L2 regularization and implemented as improved optimizer. This work introduces a novel implementation of AdamW variant as AdamW+ by further simplifying weight decay implementation in AdamW. DGA malware detection LSTM models results for Adam, AdamW and AdamW+ are evaluated on various DGA families/ groups as multiclass text classification. Proposed AdamW+ optimizer results has shown improvement in all standard performance metrics over Adam and AdamW. Analysis of outcome has shown that novel optimizer has outperformed both Adam and AdamW text classification based problems.

• Journal of the Korean Society of Radiology
• /
• v.18 no.1
• /
• pp.21-26
• /
• 2024

The one of the famous deep learning models for object detection task is you only look once version 5 (YOLOv5) framework based on the one stage architecture. In addition, YOLOv5 model indicated high performance for accurate lesion detection using the bottleneck CSP layer and skip connection function. The purpose of this study was to evaluate the performance of YOLOv5 framework according to various hyperparameters in position emission tomogrpahy (PET) phantom images. The dataset was obtained from QIN PET segmentation challenge in 500 slices. We set the bounding box to generate ground truth dataset using labelImg software. The hyperparameters for network train were applied by changing optimization function (SDG, Adam, and AdamW), activation function (SiLU, LeakyRelu, Mish, and Hardwish), and YOLOv5 model size (nano, small, large, and xlarge). The intersection over union (IOU) method was used for performance evaluation. As a results, the condition of outstanding performance is to apply AdamW, Hardwish, and nano size for optimization function, activation function and model version, respectively. In conclusion, we confirmed the usefulness of YOLOv5 network for object detection performance in nuclear medicine images.

• Journal of the Korean Society of Radiology
• /
• v.17 no.5
• /
• pp.709-715
• /
• 2023

In this study, the performance evaluation of image quality for noise reduction was implemented using the U-net deep learning architecture in computed tomography (CT) images. In order to generate input data, the Gaussian noise was applied to ground truth (GT) data, and datasets were consisted of 8:1:1 ratio of train, validation, and test sets among 1300 CT images. The Adagrad, Adam, and AdamW were used as optimizer function, and 10, 50 and 100 times for number of epochs were applied. In addition, learning rates of 0.01, 0.001, and 0.0001 were applied using the U-net deep learning model to compare the output image quality. To analyze the quantitative values, the peak signal to noise ratio (PSNR) and coefficient of variation (COV) were calculated. Based on the results, deep learning model was useful for noise reduction. We suggested that optimized hyper parameters for noise reduction in CT images were AdamW optimizer function, 100 times number of epochs and 0.0001 learning rates.

• Journal of Power Electronics
• /
• v.12 no.1
• /
• pp.33-39
• /
• 2012

This paper describes operating and capacitor voltage balancing of the modular multilevel converter. The paper focuses on sizing of the cell capacitor and establishes approximate expressions for the capacitor voltage. Simulations and experiments results obtained from three-level modular converter are used to demonstrate its viability in medium voltage applications. It is shown that the modular converter can operate over the full modulation index linear range independent of load power factor.

• 한국정보디스플레이학회:학술대회논문집
• /
• 2006.08a
• /
• pp.637-640
• /
• 2006

DuPont Teijin Films (DTF) have developed engineered substrates specifically for the flexible electronics market. $Teonex^{(R)}Q65$ is a biaxially oriented crystalline polyester with a tailored surface and it is emerging as a competitive material for the base substrate in OLED displays and active matrix backplanes. Given the dimensional reproducibility requirements in the display applications, uncontrolled moisture absorption during the processing cycle could potentially be far more significant than the inherent shrinkage of the base substrate. Understanding these effects and optimising the processing steps involved in device manufacture will be critical to achieving the ultimate performance that can be achieved with the base substrate.

• Rapid Communication in Photoscience
• /
• v.1 no.2
• /
• pp.58-58
• /
• 2012

• Smart Structures and Systems
• /
• v.19 no.2
• /
• pp.225-236
• /
• 2017

Railroad bridges in North America are an integral but aging part of the railroad network and are typically only monitored using visual inspections. When quantitative information is required for assessment, railroads often monitor bridges using accelerometers. However without a sensor to directly measure displacements, it is difficult to interpret these results as they relate to bridge performance. Digital Image Correlation (DIC) is a non-contact sensor technology capable of directly measuring the displacement of any visible bridge component. In this research, a railroad bridge was monitored under load using DIC and accelerometers. DIC measurements are directly compared to serviceability limits and it is observed that the bridge is compliant. The accelerometer data is also used to calculate displacements which are compared to the DIC measurements to assess the accuracy of the accelerometer measurements. These measurements compared well for zero-mean lateral data, providing measurement redundancy and validation. The lateral displacements from both the accelerometers and DIC at the supports were then used to determine the source of lateral displacements within the support system.

• Journal of the Korean earth science society
• /
• v.44 no.4
• /
• pp.275-290
• /
• 2023

Structural inversion refers to the reverse reactivation of extensional faults that influence basin shortening accommodated by contractional faults or folds. On the Korean peninsula, Miocene inversion structures have been found, but the Cretaceous rocks on Geoje Island may have undergone inversion as early as the Upper Cretaceous. To evaluate the structural inversion on Geoje Island, located on the eastern side of South Korea, and to determine the effects of preexisting weakness zones, field-based geometric and kinematic analyses of faults were performed. The lithology of Geoje Island is dominated by hornfelsified shale, siltstone, and sandstone in the Upper-Cretaceous Seongpori formation. NE and NW-oblique normal faults, conjugate strike-slip (NW-sinistral transpressional and E-W-dextral transtensional) faults, and NE-dextral transpressional faults are the most prominent structural features in Geoje Island. Structural inversion on Geoje Island was evidenced by the sinistral and dextral transpressional reactivation of the NW and NE-trending oblique normal faults respectively, under WNW-ESE/NW-SE compression, which was the orientation of the compressive stress during the Late Cretaceous to Early Cenozoic.

• Korean Journal of Logic
• /
• v.12 no.2
• /
• pp.31-58
• /
• 2009

This paper defends Stalnaker's theory of indicative conditionals. His theory consists of selection functions and pragmatic constraints. The selection function takes a certain possible world(W) and a proposition(A) to yield a possilble world that is similar to W and in which A is true. And the pragmatic constraints plays role to make selection functions apply just to indicative conditionals. According to Stalnaker, as indicative conditionals has strong truth-value, uncontested principle always holds but passage principle does not always hold. However, his theory can explain why passage principle sometimes holds by means of pragmatic constraints. Also, this paper argues that Stalnaker's theory is the most acceptable one among others, by replying to criticisms suggested by Adamsians and the problem raised by Gibbard and other criticisms.

• ALGAE
• /
• v.24 no.1
• /
• pp.1-29
• /
• 2009

A synthesis of early exploration and the discovery of marine algae in the upper North Pacific and Bering Sea is presented covering the period from the late 1730s to around 1900. Information is provided about these early efforts to gather natural objects, including seaweeds, and names of these algae are enumerated. The first collections of marine algae in this broad region were those made by steller and Kracheninnkov from the Kamchatka Peninsula,Russia,during the Second Kamchatkan Expedition (1735-1742) and were described by Gmelin (1768). The first known algal collections in Alaska were those made byMerck in his 1790-1791 visits to Unalaska Island during the Billings expedition (1785-1794). British-sponsored expeditions for commercial purposes and for exploration and dis-covery allowed surgeon-naturallist Archibald Menzies to garher seaweeds that Dawson Turner and others worked up back in Europe. Several of the Russian Expeditions during the first half of the 18'!' century had naturalists aboard. the first Russian circumnavigation of the globe (1803-1806), with the ships 'Nadeshda' and 'Neva,' under the com-mand of Capt. Adam von Krusenstern had naturalists Langsdorff, Tilesius, and Horner, all of whom collected sea-weeds. The naturalist Adelbert Chanmisso accompanied the Romanzof Expedition (1815-1818) on the Russian vessel 'Rurik' under the command of Otto von Kotzebue and made collections of algae in the Aleutians as well as in the Kurils and Kamchatka. The Lutke expedition of 1826-1829 consisted of thw ships. Feodor Lutke was in command of the 'Seniavin' with K.H. Mertens aboard as physician-naturalist, and the 'Moller' was under the command of staniukovich accompanied by the naturalist G. Kastalsky. The first American-sponsored scientific expedition (1838-1842) was that commanded by Charles Wilkes, and the algae that were collected were worked up by J.W. Bailey and W.H. Harvey. The Russian naturalist Ilya Voznesenskii spent the period 1839-1849 in Russian Americ (Alaska and northern California) energetically traveling and making numerous collections of natural objects as well as ethno-graphic artefact. His algae were described by F.j. Ruprecht back in St. petersbung. The Swedish scientific vessel, the'Vega' (1878-1880), was under the command of Nordenskiold. The naturalist F.R. Kjellman made algal collections from Port Clarence, Alaska, as well as from bering Island and St. Lawrence Island in the Bering sea. The Harriman Alaskan Expedition in the summer of 1899, with the ship 'George W. Elder,' was sponsored by railroad magnate E.H. Harriman of New York City and had several scientific personnel aborad, including the phycologist De Alton Saunders. Algae were collected in Alaska and Washington. During the same summer of 1899 a scientific expedition organized by the University of California and including W.L. Jepson, L.E. Hunt, A.A Lawson, and W.A. Setchell as participants also visited Alaska and made collections of alage from various locations.