중국어 텍스트 분류 작업의 개선을 위한 WWMBERT 기반 방식

왕흠원 ( Xinyuan Wang ),조인휘 ( Inwhee Joe ) 한국정보처리학회 2021 한국정보처리학회 학술대회논문집 Vol.28 No.1

In the NLP field, the pre-training model BERT launched by the Google team in 2018 has shown amazing results in various tasks in the NLP field. Subsequently, many variant models have been derived based on the original BERT, such as RoBERTa, ERNIEBERT and so on. In this paper, the WWMBERT (Whole Word Masking BERT) model suitable for Chinese text tasks was used as the baseline model of our experiment. The experiment is mainly for "Text-level Chinese text classification tasks" are improved, which mainly combines Tapt (Task-Adaptive Pretraining) and "Multi-Sample Dropout method" to improve the model, and compare the experimental results, experimental data sets and model scoring standards Both are consistent with the official WWMBERT model using Accuracy as the scoring standard. The official WWMBERT model uses the maximum and average values of multiple experimental results as the experimental scores. The development set was 97.70% (97.50%) on the "text-level Chinese text classification task". and 97.70% (97.50%) of the test set. After comparing the results of the experiments in this paper, the development set increased by 0.35% (0.5%) and the test set increased by 0.31% (0.48%). The original baseline model has been significantly improved.

The role of Jacalin-related lectin gene AOL_s00083g511 in the development and pathogenicity of the nematophagous fungus Arthrobotrys oligospora

Dong Xinyuan,Si Jiali,Zhang Guanghui,Shen Zhen,Zhang Li,Sheng Kangliang,Wang Jingmin,Kong Xiaowei,Zha Xiangdong,Wang Yongzhong 한국미생물학회 2021 The journal of microbiology Vol.59 No.8

Arthrobotrys oligospora is a model species of nematophagous fungi and has great potential for the biological control of nematode diseases. Lectin is a protein that binds to carbohydrates and their complexes with high specificity, which mediates recognition events in various physiological and pathological processes. This study aimed to investigate the role of the Jacalin-related lectin (JRL) gene, AOL_s00083g511, in A. oligospora development. Through a homology recombination approach, we obtained the AOL_s00083g511 knockout mutant strain (Δg511). Next, the biological characteristics of the Δg511 mutant strain, including growth rate, conidia germination rate, adaptation to environmental stresses, and nematocidal activity, were compared with those of the wild-type (WT) strain. The results showed that the JRL gene AOL_ s00083g511 did not affect fungal growth, conidia germination, 3D-trap formation, and the ability of A. oligospora to prey on nematodes significantly. We speculate that this phenomenon may be caused by a loss of the key β1–β2 loops in the AOL_ s00083g511-encoded JRL domain and an intrinsic genetic compensation of AOL_s00083g511 in this fungus. The growth rates of both strains on high salt or surfactant media were similar; however, in the strong oxidation medium, the growth rate of the Δg511 mutant was significantly lower than that of the WT strain, indicating that AOL_s00083g511 might play a role in oxidative stress resistance. These findings provide a basis for further analysis of the related functions of the JRL gene in A. oligospora and their potential roles in the biological control of nematodes in the future.

Photocatalytic Degradation of Rhodamine B via Fe-g-C3N4 Activated Sulfate Radical-Based Advanced Oxidation Processes and the Synergistic Mechanism

Zhan Xinyuan,She Zhixiang,Yue Zhengbo,Hu Fupeng,Wang Guangcheng,Wang Shaoping,Li Wei,Liu Bing,Wang Jin 한국화학공학회 2024 Korean Journal of Chemical Engineering Vol.41 No.1

Fe-doped g-C 3 N 4 has been proven to have the potential of visible light photocatalysis, but its catalytic activity of peroxymonosulfate (PMS) is neglected. Herein, the complex advanced oxidation processes of Fe-g-C 3 N 4 mediated PMS and visible light photocatalysis was developed, named as Vis/Fe-g-C 3 N 4 /PMS system, whose eff ects and synergistic mechanism for decomposing Rhodamine B (RhB) was evaluated. The coupling of sulfate radicals and photocatalysis for RhB degradation showed that the synergistic effi ciency ( η Syn ) and factor ( S c ) were 20.1% and 4.82, respectively, with a degradation effi ciency of 99.8%. Iron species dispersed on g-C 3 N 4 provided active sites for PMS activation to generate sulfate radicals, simultaneously reduced the forbidden band, and separated the photo-generated charges of g-C 3 N 4 . h + , SO 4 · ‾ and 1 O 2 were the main active species, and the increase of 1 O 2 was the cause of the synergistic eff ect. The possible degradation path of RhB by this coupling system was proposed. Our fi ndings prove that Vis/Fe-g-C 3 N 4 /PMS system has a great potential to decompose dye wastewater, and also to be an environmental remediation perspective.

Surface Modification of Flax Fibers with Isocyanate and Its Effects on Fiber/Epoxy Interfacial Properties

Wanru Wang,Rao Fu,Qi Deng,Xinyuan Wang,Yujie Wang,Zhao Zhang,Guijun Xian 한국섬유공학회 2020 Fibers and polymers Vol.21 No.12

Natural fiber reinforced polymer composite has been widely used in various industry fields. Natural fibertreatment can effectively improve its mechanical and durability properties, and expands its applications. In the present study,isocyanate was proposed to treat flax fibers in order to reduce the hydrophilic properties of the fibers, and to enhance thebonding of flax fiber to epoxy matrix. The isocyanate treated fabric was evaluated with fourier transform infra-red (FTIR),scanning electron microscopy (SEM) and water uptake. The effects of the fiber treatment on the mechanical properties of flaxfabric reinforced epoxy (FFRP) plates were investigated. FTIR and SEM analysis indicated that isocyanate reacts with flaxfiber, forming a thin polymer layer on the fiber surface. The water absorption test showed that isocyanate treatment decreasedthe water uptake of the flax fabric by 4-18 times compared to the control fabric. The flexural and tensile strength of FFRPwith isocyanate solution treated fabrics were enhanced more than 20 %, attributed to the improved adhesion of fiber to epoxy. Based on the above results, the treatment method of flax fiber with isocyanate is considered as an effective approach toimprove both hydrothermal ageing resistance and mechanical properties of FFRPs.

A Novel Image Superresolution Reconstruction Algorithm Based on Sparse Representation

Aili Wang,Xinyuan Wang,Yuji Iwahori,Yuan Feng,Na Jiang 보안공학연구지원센터 2015 International Journal of Signal Processing, Image Vol.8 No.6

Superresolution image reconstruction technique uses single or a series of low-resolution images to reconstruct a high resolution image without changing the hardware devices, while improving image quality and the spatial resolution of the image. High resolution means the image with a higher pixel density, can provide more details. In this paper, a novel image superresolution algorithm based on sparse representation is studied. During over-complete dictionary of the training phase, the proposed method improves two aspects including feature extraction and dimension reduction. In the feature extraction process, combining the second derivative with the gradient direction, we construct a new descent direction to improve gradient method. The convergence speed of the new algorithm is faster than the gradient method and can get better results. Then improved two-dimensional Principal Component Analysis (2DPCA) algorithm is used to reduce the dimension, it could eliminate the correlation of the image lines and column. Experiment results show that this method of image reconstruction is better and faster for high resolution image reconstruction.

Thermo-mechanical Coupling Modeling of Active Phased Array Antennas

Xinyuan Wei,Enming Miao,Yangyang Chen,Wei Wang 한국정밀공학회 2019 International Journal of Precision Engineering and Vol.20 No.11

The large thermal deformation of the panel in active phased array antennas (APAAs) caused by the thermal power consumption of the internal devices and the change in the ambient temperature in actual operating conditions will result in a considerable decline in the electromagnetic (EM) performance of radar. Experimental research indicates that the thermal deformation of the panel is affected not only by temperature but also by the constraints caused by the installation. Thus, this paper proposes a thermo-mechanical coupling modeling method for an APAA considering the panel temperature and constraints. The coupling model can accurately predict the panel thermal deformation, thereby providing a basis for the compensation of the EM performance of radar. The thermal deformation of the panel is decomposed into a thermal expansion and an offset. Models of the thermal expansion with respect to temperature and the offset with respect to the constraint force are established. The final thermo-mechanical coupling model is obtained by coupling the two models. Experiments verify that the prediction accuracy of the coupling model is 86.3%, which is 21.0% higher than that of a finite element software simulation. The content of this research has practical engineering importance for improving the overall performance of APAAs

Study of generalized electric spring modeling based on Hamilton’s principle and its stability

Xiaohu Wang,Chaohui Zhao,Xinyuan Chen,Zhun Huang 전력전자학회 2024 JOURNAL OF POWER ELECTRONICS Vol.24 No.5

The proposed generalized electric spring (G-ES) topology effectively reduces the redundancy caused by the parallel structure of multiple electric springs in a microgrid system. A modeling method for the G-ES is urgently needed to accurately determine the G-ES parameters and to evaluate its transient operation characteristics. The correspondence between Hamilton’s principle in mechanics and electricity is introduced, and the feasibility of Hamilton’s principle under the smart load topology is verified. The modeling method of the G-ES under Hamilton’s principle is discussed, followed by the application of repetitive control strategies on the G-ES. The filter parameters are designed using the normalization method and a simplified parameter design using repetitive controllers was applied. This article introduces the impedance analysis method into the generalized electric spring to analyze its stability under weak network conditions. This work also derives the stability judgment criteria for generalized electric springs. Finally, the feasibility of modeling the G-ES based on the Hamilton’s principle was verified using MATLAB and a DSP hardware system. In addition, the consistency between the G-ES and a single intelligent load was verified under the Hamilton’s modeling principle. The G-ES stability and dynamic performance under this modeling method meet industry requirements.

Psychological Resilience of Second-Pregnancy Women in China: A Cross-sectional Study of Influencing Factors

Xiaohuan Jin,Xinyuan Xu,Junyan Qiu,Zexun Xu,Lixue Sun,Zhilin Wang,Ling Shan 한국간호과학회 2021 Asian Nursing Research Vol.15 No.2

Purpose The aim of the study was to evaluate the status of psychological resilience among women in their second pregnancy and to investigate the possible influencing factors. Methods A total of 275 women in their second pregnancy and who met the criteria were surveyed from two public hospitals in China from July 2018 to January 2019. The instruments included the General Self-designed Questionnaire, Connor–Davidson Resilience Scale, Social Support Rate Scale, and 36-item Pregnancy Stress Rating Scale. Results The total psychological resilience score of second-pregnancy women was relatively low. Multivariate regression analysis identified five factors associated with psychological resilience: intimacy with husbands, social support utilization, gender of the first child, high-risk pregnancy of the first child, and the stress caused by worrying about the health and safety of the mother and fetus. Conclusion Women in their second pregnancy represent a unique population, and their low psychological resilience score deserves attention. Identification of factors contributing to decreased psychological resilience may enable us to design prevention and intervention strategies and to deliver specific psychological supports to pregnant women at high risk of developing negative psychology.

Decrease of Oxygen Vacancy by Zn-Doped for Improving Solar-Blind Photoelectric Performance in β-Ga2O3 Thin Films

Daoyou Guo,Xinyuan Qin,Ming Lv,Haoze Shi,Yuanli Su,Guosheng Yao,Shunli Wang,Chaorong Li,Peigang Li,Weihua Tang 대한금속·재료학회 2017 ELECTRONIC MATERIALS LETTERS Vol.13 No.6

Highly (201) oriented Zn-doped β-Ga2O3 thin films with differentdopant concentrations were grown on (0001) sapphire substrates byradio frequency magnetron sputtering. With the increase of Zn dopantconcentration, the crystal lattice expands, the energy band gap shrinks,and the oxygen vacancy concentration decreases. Both the metalsemiconductor metal (MSM) structure photodetectors based on thepure and Zn-doped β-Ga2O3 thin films exhibit solar blind UVphotoelectric property. Compared to the pure β-Ga2O3 photodetector,the Zn-doped one exhibits a lower dark current, a higher photo/darkcurrent ratio, a faster photoresponse speed, which can be attributed tothe decreases of oxygen vacancy concentration.

A Novel Feature Detection Algorithm Based on Improved 2DPCA- SIFT

AiliWang,Yangyang Zhao,Jiaying Zhao,Yuji Iwahori,Xinyuan Wang 보안공학연구지원센터 2015 International Journal of u- and e- Service, Scienc Vol.8 No.6

Stable local feature and representation is a fundamental component of many image registration, 3D reconstruction and object recognition algorithms. SIFT is a good descriptor that encodes the salient aspects of the image gradient in the feature point’s neighborhood. This paper improved SIFT- based local image descriptor and proposed a SIFT feature matching algorithm based on improved 2DPCA which can eliminate both rows and columns of relevance. Experimental results show that improved 2DPCA-SIFT algorithm is relatively stable, accurate and fast.