• The Journal of Bigdata
• /
• v.7 no.2
• /
• pp.11-29
• /
• 2022

Pre-trained language models are the most important and widely used tools in natural language processing tasks. Since those have been pre-trained for a large amount of corpus, high performance can be expected even with fine-tuning learning using a small number of data. Since the elements necessary for implementation, such as a pre-trained tokenizer and a deep learning model including pre-trained weights, are distributed together, the cost and period of natural language processing has been greatly reduced. Transformer variants are the most representative pre-trained language models that provide these advantages. Those are being actively used in other fields such as computer vision and audio applications. In order to make it easier for researchers to understand the pre-trained language model and apply it to natural language processing tasks, this paper describes the definition of the language model and the pre-learning language model, and discusses the development process of the pre-trained language model and especially representative Transformer variants.

• Journal of Broadcast Engineering
• /
• v.23 no.1
• /
• pp.162-165
• /
• 2018

In this paper, we propose a pre-processing procedure to obtain scene change information using spatial down-scaled input image for efficient encoding of super-high resolution image and propose a reconstruction of reference picture list in inter-picture prediction using this information. The experimental results show that the proposed method improves the BD-Rate by 0.44% and reduces encoding time by 12.46% when compared to HM 16.12.

• Food Science of Animal Resources
• /
• v.38 no.5
• /
• pp.970-980
• /
• 2018

We evaluated the effects of pre-cooking methods on the quality of reheated marinated pork loin. Frozen marinated pork loins cooked using various methods (boiling, grilling, pan frying, infrared cooking, and superheated steam cooking) were reheated in a microwave, and their pH, color, cooking loss, re-heating loss, total loss, thiobarbituric acid reactive substance (TBARS) value, sensory properties, and shear force were determined. Although all parameters varied with different cooking methods, lightness values and TBARS values showed the tendency to decrease and increase, respectively, after reheating. Superheated steam-cooked samples showed the lowest values of cooking loss, total loss, TBARS value, and shear force (p<0.05) and the highest lightness, redness, and yellowssness values and juiciness, chewiness, and overall acceptability scores (p<0.05). These results show that pre-cooking with superheated steam maintains the quality characteristics of marinated pork loin upon reheating. Therefore, pre-cooking with superheated steam may be beneficial for the commercial distribution of frozen cooked marinated pork loin.

• The Journal of Information Technology
• /
• v.7 no.2
• /
• pp.43-54
• /
• 2004

Since Dr. John J. Hopfield has proposed the HOpfield network, it has been widely applied to the pattern recognition and the routing optimization. The method of Jian-Hua Li improved efficiency of Hopfield network which input pattern's weights are regenerated by SVD(singluar value decomposition). This paper deals with Li's Hopfield Network by linear pre-processing. Linear pre-processing is used for increasing orthogonality of input pattern set. Two methods of pre-processing are used, Hadamard method and random method. In manner of success rate, radom method improves maximum 30 percent than the original and hadamard method improves maximum 15 percent. In manner of success time, random method decreases maximum 5 iterations and hadamard method decreases maximum 2.5 iterations.

• Journal of Welding and Joining
• /
• v.34 no.2
• /
• pp.11-15
• /
• 2016

FE analyses for weldment of ship structure are required for various reasons such as stress concentration for bead tow, residual stress and distortion after welding, and hydrogen diffusion for prediction of low temperature crack. These analyses should be done by solid element modeling, but most of ship structures are modeled by shell element. If we are able to make solid element in the shell element FE modeling it is easily to solve the requirement for solid elements in weld analysis of large ship structures. As the nodes of solid element cannot take moments from nodes of shell element, these two kinds of element cannot be used in one model by conventional modeling. The PSCM (Perpendicular shell coupling method) can connect shell to solid. This method uses dummy perpendicular shell element for transferring moment from shell to solid. The target of this study is to develop a FE pre-processing system applicable at welding at ship structure by using PSCM. We also suggested glue-contact technique for controlling element numbers and element qualities and applied it between PSCM and solid element in automatic pre-processing system. The FE weldment modeling through developed pre-processing system will have rational stiffness of adjacent regions. Then FE results can be more reliable when turn-over of ship-block with semi-welded state or ECA (Engineering critical assessment) of weldment in a ship-block are analyzed.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.15 no.5
• /
• pp.1164-1170
• /
• 2011

In this paper, we propose accurate lattice extraction of elemental image array and pre-processing methods in computational integral imaging. Pre-processing methods remove distortions and noises of the image. Such distortions occurred in pickup systems are rotational errors. Distortions will degrade the resolution of reconstructed images. To overcome this problem, we propose our methods for extraction of elemental image array and pre-processing methods. Also, we describe that distortions affect the high quality reconstruction. Optical and computational experiments indicate that reconstructed images applied our methods is better than reconstructed images unapplied our methods.

• Journal of Sensor Science and Technology
• /
• v.17 no.6
• /
• pp.406-413
• /
• 2008

We have been implemented simple statistical pattern recognition methods for harmful gases classification using gas sensors array fabricated by MEMS (Micro Electro Mechanical System) technology. The performance of pattern recognition method as a gas classifier is highly dependent on the choice of pre-processing techniques for sensor and sensors array signals and optimal classification algorithms among the various classification techniques. We carried out pre-processing for each sensor's signal as well as sensors array signals to extract features for each gas. We adapted simple statistical pattern recognition algorithms, which were PCA (Principal Component Analysis) for visualization of patterns clustering and MLR (Multi-Linear Regression) for real-time system implementation, to classify harmful gases. Experimental results of adapted pattern recognition methods with pre-processing techniques have been shown good clustering performance and expected easy implementation for real-time sensing system.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.20 no.4
• /
• pp.109-115
• /
• 2010

Differential Power Analysis (DPA) is well known as one of efficient physical side-channel attack methods using leakage power consumption traces. However, since the power traces usually include the components irrelevant to the encryption, the efficiency of the DPA attack may be degraded. To enhance the performance of DPA, we introduce a pre-processing technique which extracts the encryption-related parts from the measured power consumption signals. Experimental results show that the DPA attack with the use of the proposed pre-processing method detects correct cipher keys with much smaller number of signals compared to that of the conventional DPA attack.

• Journal of Multimedia Information System
• /
• v.8 no.2
• /
• pp.79-84
• /
• 2021

Detecting brain tumors of different sizes is a challenging task. This study aimed to identify brain tumors using detection algorithms. Most studies in this area use segmentation; however, we utilized detection owing to its advantages. Data were obtained from 64 patients and 11,200 MR images. The deep learning model used was RetinaNet, which is based on ResNet152. The model learned three different types of pre-processing images: normal, general histogram equalization, and contrast-limited adaptive histogram equalization (CLAHE). The three types of images were compared to determine the pre-processing technique that exhibits the best performance in the deep learning algorithms. During pre-processing, we converted the MR images from DICOM to JPG format. Additionally, we regulated the window level and width. The model compared the pre-processed images to determine which images showed adequate performance; CLAHE showed the best performance, with a sensitivity of 81.79%. The RetinaNet model for detecting brain tumors through deep learning algorithms demonstrated satisfactory performance in finding lesions. In future, we plan to develop a new model for improving the detection performance using well-processed data. This study lays the groundwork for future detection technologies that can help doctors find lesions more easily in clinical tasks.

• Proceedings of the Korean Society of Dyers and Finishers Conference
• /
• 2009.03a
• /
• pp.43-44
• /
• 2009

Fabrics for Digital Textile Printing (DTP) are different from it of general textile printing. It is necessary to pre-treatment of chemical agents for desired quality. But this process does not correspond with simplification of DTP processing. In this research, we pre-treated of cotto is necessary to pre-treatment of chemical agents for desired quality. But this process does not correspond with simplification of DTP processing. In this research, we pre-treated of cotton fabric for DTP by atmosphere plasma treatment and we understood that pre-treatment of fabric by atmosphere plasma treatment was more simple DTP process.