GIST Scholar
Not a member yet
30271 research outputs found
Sort by
Soil and Vegetation Characteristics Associated with the Invasion of Humulus japonicus in Riparian Zones
No full textRiparian zones are frequently exposed to floods and artificial disturbances, promoting the invasion of aggressive native plant species. Humulus japonicus is particularly dominant in Korean riparian areas, where its vining habit enables it to overgrow surrounding vegetation. Despite being designated as an ecosystem-threatening species by the Ministry of Environment, ecological insights into its invasive success remain limited. We compared soil and vegetation characteristics between sites invaded and uninvaded by H. japonicus to identify habitat preferences. Additionally, we assessed sex ratios and seed banks, which are critical for the maintenance of dioecious annual plant populations like H. japonicus. Uninvaded sites had significantly higher coverage of perennial plants, and the coverage of H. japonicus was negatively correlated with that of perennial plants across invaded sites. In contrast, the coverage of annual and biennial species was positively associated with the presence of H. japonicus. Among soil characteristics identified by canonical correspondence analysis as explanatory variables for species composition, soil water content was higher in uninvaded sites than in invaded sites. These results suggest that the establishment of H. japonicus is associated with the structure of riparian vegetation and soil water content, although the causal relationships among those remain unclear. No viable H. japonicus seeds were found in uninvaded soils, and only a small number were detected in invaded sites, indicating that seed dispersal from upstream may play a critical role in maintaining population size. The biotic and abiotic factors identified in this study warrant further investigation to clarify their causal roles in the invasion of H. japonicus through manipulative experiments. © The Author(s), under exclusive licence to Korean Society of Plant Biologist 2025.FALSEsciescopuskc
A High Performance Polarization Sensitive Photodetector Based on Remote Doped van der Waals Heterostructure
No full textPhotodetector, which converts photon energy of light to the electrical signal, is widely applied to applications such as LiDAR, airline, terrain, astronomy fields which results in increasing demand of high performance photodetector. Two-dimensional (2D) material based photodetector has received more interest thanks to its novel optoelectronic properties such as high quantum efficiency at low thickness, broadband photodetection. However, 2D material has low light absorption properties, which is closely related to the single atomic thickness. Doping is widely applied to enhance photoresponsivity, but has a trade off with photoresponse time due to charged impurity scattering by introduced dopants. Here, we applied remote modulation doping method to ReS2 photodetector by simple insertion of monolayer hexagonal boron nitride between doping solution and channel material. With remote modulation doping, we achieves simultaneous enhancement of photoresponsivity and temporal response time. In addition, with intrinsically anisotropic crystal structure of ReS2, polarization sensitive photoresponse, which provides efficient target recognition, enhanced display and imaging, is confirmed with ReS2 photodetector and also demonstrated that doping does not have negative effect to the polarization sensitivity. The proposed high performance and polarization sensitive photodetection technology promotes a new way for two-dimensional material based optoelectronic devices.MasterAbstract ․․․․․․․․․․․․․․․․․․․․ i
Introduction ․․․․․․․․․․․․․․․․․․․․ P. 1
Results and Discussion ․․․․․․․․․․․․․․․․․․․․ P.2~ 9
Summary ․․․․․․․․․․․․․․․․․․․․ P.10~ 11
Experimental Section ․․․․․․․․․․․․․․․․․․․․ P.12~ 13
Reference ․․․․․․․․․․․․․․․․․․․․ P.14~ 15
Acknowledgement․․․․․․․․․․․․․․․․․․․․ P. 16
Curriculum Vitae ․․․․․․․․․․․․․․․․․․․․ P. 1
Study on Unsupervised Learning and Cyber Threat Detection in Industrial Control Systems Woo-Hyun Choi Gwangju Institute of Science and Technology
No full text산업 제어 시스템(ICS)은 제조 및 에너지와 같은 분야의 중요 인프라를 관리하는 데 필수적입니다. 이러한 시스템들의 네트워크 연결성이 증가함에 따라 사이버 보안 위험 에 대한 취약성이 높아졌습니다. 격리된 환경을 위해 설계된 ICS는 이제 중요한 위협에 노출되어 있으며, 이는 중요 인프라에 물리적 손상을 초래한 Stuxnet 공격과 같은 사례 에서확인되었습니다.전통적인 IT시스템과달리 ICS는 물리적 프로세스를 제어하므로 사이버 공격이 실제 세계에 구체적인 영향을 미칠 수 있습니다. ICS 보안의 주요 과 제는 다양한 장치와 독점 프로토콜로 구성된 이러한 시스템의 복잡성과 이질성입니다. 지속적인 운영 요구로인해 보안 업데이트 기회가 제한되어 기존 IT보안조치의 효과가 감소됩니다. 그결과, 실시간으로 알려지지 않은 위협과 이상을 탐지하기 위해 머신러닝과 비지도학습기술을 포함한 고급 접근 방식을 탐구하는 데 관심이 증가하고 있습니다. 이 논문은 두 부분으로 구성되어 있습니다. 첫 번째 부분은 비지도 기계 학습을 사용한 ICS의 이상 탐지를 검토합니다. 이 연구는 사전에 레이블이 지정된 데이터 없이 이상 행동을 식별하기 위한 복합 오토인코더 모델을 조사합니다. HIL-based Augmented ICS(HAI)의 데이터셋을 활용하여 값과 시간 모두와 관련된 이상을 탐지하는 모델의 능력을 분석합니다. 이 접근 방식은 ICS 환경 에서 시스템 신뢰성과 운영 효율성을 향상시키기 위한 지속적인 노력에 기여하는 것을 목표로 합니다. 두 번째 부분에서는 MITRE ATT&CK 프레임워크와 함께 ICS 트래픽의 이상 탐지를 위한 영과잉 포아송(ZIP) 기반 GRU 학습 모델을 탐구합니다. 모델의 성능은 ’Stuxnet’ 과 ’Industroyer’라는 두가지 주요 사이버 공격 시나리오 시뮬레이션을 통해 평가 되었습니다. 탐지된 이상을 MITRE ATT&CK 프레임워크에 매핑함으로써, 이 연구는 이러한 공격에 대한 정보에 입각한 대응 전략 개발에 기여하고자 합니다. 이 연구는 ICS 보안의 지속적인 과제를 다루며,진화하는 사이버위협으로부터 이러한 중요시스템을 보호하기 위한 잠재적 접근 방식을 연구합니다.|Industrial Control Systems (ICS) are essential for managing critical infrastructure in sectors such as manufacturing and energy. The increasing connectivity of these systems to networks has heightened their vulnerability to cybersecurity risks. Originally designed for isolated environments, ICS are now exposed to significant threats, as evidenced by incidents like the Stuxnet attack, which resulted in physical damage to critical infrastructure. Unlike traditional IT systems, ICS control physical processes, meaning cyberattacks can have tangible, real-world impacts. A significant challenge in ICS security is the complexity and heterogeneity of these systems, which comprise diverse devices and proprietary protocols. The requirement for continuous operation limits opportunities for security updates, reducing the effectiveness of conventional IT security measures. As a result, there is growing interest in exploring advanced approaches, including machine learning and unsupervised learning techniques, for real- time detection of unknown threats and anomalies. This dissertation examines two key topics. The first examines anomaly detection in ICS using unsupervised machine learning. The study investigates a composite autoencoder model for identifying anomalous behavior without pre-labeled data. Utilizing a dataset from HIL-based Augmented ICS (HAI), the research analyzes the model’s capacity to detect anomalies related to both value and time. This approach aims to contribute to the ongoing efforts to enhance system reliability and operational efficiency in ICS environments. The second focuses on a Zero Inflated Poisson (ZIP) based GRU Learning model for anomaly detection in ICS traffic, in conjunction with the MITRE ATT&CK framework. The model’s performance was evaluated through simulations of two major cyberattack scenarios, Stuxnet and Industroyer. By mapping detected anomalies to the MITRE ATT&CK framework, the study seeks to contribute to the development of more in- formed response strategies for such attacks. This research addresses the ongoing challenges in ICS security and studies potential approaches to enhance the protection of these critical systems against evolving cyber threats.DoctorAbstract (English) i
Abstract (Korean) iii
List of Contents v
List of Tables vii
List of Figures ix
1 Introduction 1
1.1 Research Motivation 2
1.2 Contributions 8
1.3 Organization 10
2 Related Work 12
2.1 Anomaly Detection and Identification in Complex Industrial Control
Systems 12
2.1.1 Anomaly Detection in Industrial Control Systems 13
2.1.2 Recent Approaches to the Study of Anomaly Detection in Industrial Control System 14
2.2 GRU-Poisson Model 16
2.2.1 MITRE ATT&CK Framework for ICS 16
2.2.2 ICS Security and Protection Mechanisms 19
3 Unsupervised Learning Approach for Anomaly Detection in Indus-
trial Control Systems 24
3.1 Proposed Approach 25
3.2 HAI Dataset 25
3.2.1 Data Preparation 28
3.2.2 Evaluation Metrics 36
3.3 Experiment and Evaluation 37
3.3.1 Classification and Confusion Matrix 37
3.3.2 Abnormal Behavior Detection Experiment 39
3.4 Summary 43
4 Detecting Cybersecurity Threats for Industrial Control Systems using Machine Learning 44
4.1 Motivation, Problem Description, and Key Contributions 45
4.2 System Model 45
4.2.1 Raw traffic dataset 45
4.2.2 Aligning data with MITRE ATT&CK Framework 49
4.2.3 Zero-Inflated Poisson 51
4.2.4 GRU based Detection Model 52
4.2.5 Model Performance Comparison 57
4.3 Implementation 59
4.4 Experiment and Result and Analysis 61
4.4.1 Stuxnet 63
4.4.2 Industroyer 69
4.5 Summary 76
5 Conclusion 78
References 81
A. Abbreviations 92
Acknowledgements 9
Intensity correlations in scattered vector vortex beams for determining the topological charge
No full textWe propose a novel scheme for measuring the topological charge (TC) of the vector vortex beam (VVB) using polarization speckles size of the 2D auto-correlation function of scattered vector vortices. We have generated the non-separable state having singularities in both orbital angular momentum (OAM) and polarization. This non-separability is verified by measuring the degree of polarization (DOP) which can be calculated by using Stokes parameters. The Stokes parameters can be measured by observing the intensity distribution of the beam when projected onto different polarization states. The divergence of polarization speckle size of the auto-correlation function was measured as a function of topological charge of the vector vortex beams. These results can be used to generate a security key for data authentication and data encryption. © 2025 Elsevier B.V., All rights reserved.TRUEsciescopu
Role of ionization potential depression for generation of strongly coupled plasmas in high-pressure supercritical fluids
No full textThe photosphere, one of the Sun’s inner layers, and white dwarfs exist as plasmas with high electron density and relatively low temperature, satisfying the strongly coupled condition that the Coulomb coupling parameter exceeds unity. While strongly coupled plasmas are prevalent throughout space, it is difficult to produce and sustain such states in the laboratory with a sufficiently long lifetime for investigation of basic transport processes. To surmount these obstacles, we have developed a laser-produced plasma experiment in supercritical fluids of He, Ar, and Kr. For helium at 100 bar, a nanosecond laser pulse generates a strongly coupled plasma with an electron density, estimated from ionization potential depression (IPD) modeling, of approximately, and a temperature of about 1 eV, corresponding to a Coulomb coupling parameter of 4. Systematic experiments with other species show that the pivotal factor for achieving strongly coupled plasmas is the degree of IPD. This study shows the potential to facilitate precise measurements of thermodynamic transport parameters and equation of state for dense plasma states. © The Author(s) 2025.TRUEsciescopu
Tailoring orbital angular momentum entanglement through inversed design based on gradient descent
No full textHigh-dimensional quantum entanglement is an important resource for fundamental research and quantum technological application. Recently, the orbital angular momentum (OAM) of light has received much attention to realize high-dimensional quantum systems, and generating and manipulating the OAM entanglement has become an important topic. Here, we propose an inverse design technique to manipulate the OAM entanglement of the photon pair generated through spontaneous parametric down-conversion. The proposed technique directly tailors the OAM states of entangled photon pairs by shaping the incident pump beam. In particular, we construct a physics-informed model by leveraging the backpropagation algorithm and adhering to the overlap integral relation. Our model automatically optimizes diffractive masks, allowing the structured pump to quantitatively modulate the OAM distribution. The generation of maximally entangled states in different dimensions is demonstrated numerically and experimentally. © 2025 The Author(s)TRUEsciescopu
Identify and improve the failure mechanism of contact failure in multifunction switches by developing sequential test methods
No full textContact failure and light malfunction are common problems in multifunction swit- ches. Fretting refers to surface degradation in mechanical components. Two materials pressed together by an external static load are subjected to a transverse cyclic load or various vibrations, causing wear. Numerous studies have been conducted to address the problem of chronic signal loss. However, the cause of failure in the early stages remains obscured. The field failure analysis shows that the contact pressure decreases even at low mileage (<10000 km). Meanwhile, the finite element method analysis demonstrates that the contact pressure decreases during the assembly process of multifunction switches. We validated that high-temperature-induced creep and vibration are the primary causes of failure in multifunction switches through a series of tests. Although conventional testing methods consist of unit evaluations to verify specific failure mech- anisms, such as high-temperature exposure and vibration, and cannot reproduce actual failures, the series test method can reflect actual failure phenomena. Additionally, benchmarking with competitors confirmed that the round design structure mitigates the reduction in contact pres- sure. Furthermore, we provided guidelines for improvement from quality and design perspec- tives. The most important result of this study was the identification of the failure mechanism of multifunction switches through the series test method, and creep (high temperature) has a stronger effect on the degradation of contact dimensions and pressure than vibration.FALSEsciescopuskc
“We Are Not Luddites”: Image Generative AI and a Sociotechnical Resistance among Illustrators
No full text생성형 AI가 인간의 창의력이 요구되는 다양한 예술 영역까지 진출하면서 창작자들과 개발자 및 개발기업 사이의 갈등이 고조되고 있다. 시나리오 작가, 일러스트레이터, 만화가, 음악 프로듀서 등 다양한 예술 창작 영역에서 이런 대립과 긴장을 발견할 수 있다. 특히, 미드저니와 노벨 AI 등 이미지 생성 AI가 인터넷에서 창작자들의 작품 이미지 수억 개를 학습해 인간 그림 작가와 매우 흡사한 이미지를 생성하자, 그림 작가, 그중에서도 일러스트레이터 집단은 강한 비판의 목소리를 내어 왔다. 이들은 무분별한 이미지 학습을 방해하는 프로그램들인 나이트셰이드나 글레이즈로 AI를 망가뜨리려는 공격까지 했는데, 일부 개발기업과 온라인 커뮤니티에서는 이들의 행위를 과거 러다이트 운동처럼 시대의 변화를 거스르는 무모하고 감정적 반응으로 평가하고 있다. 이 글은 이 갈등을 기술적 합리성 대 창작자의 권리 사이의 대립으로 그리는 것은 적절치 않다고 보며, 대신 기술-문화적 조합 속에서 이 갈등을 재해석하고자 한다. 먼저 일러스트레이터 집단이 공유하는 내부 규범을 분석함으로써 이 갈등의 저변에 존재하는 보다 복잡한 층위를 드러내고자 한다. 국내의 현직 일러스트레이터들을 면담하고 관련 소셜 미디어 게시물을 분석한 이 연구는 이들의 AI 비판과 실질적 공격에는 단순히 저작권 문제를 넘어 내부 규범과 실천, 직업 생태계의 특성이 있으며 나름의 정당성이 있었다고 주장한다. 기만을 기술적 목표로 삼는 개발자 규범에 맞서 독창성 존중 문화를 일궈온 일러스트레이터의 집단 규범, 일러스트레이터 작가들이 처한 노동 환경, AI 기술이 ‘그저 그런 모방’에 가까운 점 등이 갈등 원인과 전개 과정에 있었다. 이는 결국 일러스트레이터들이 왜 공격적 대응 전략을 채택할 수밖에 없었는지도 설명할 수 있다. 결론에서 이 글은 일러스트레이터들의 저항은 예의 없는 기술에 존중을 요구하는 일종의 ‘사회기술적 저항’이었다고 평가하고 지금의 한계 있는 기술은 예술 노동의 계층화를 앞으로 심화시킬 것이라고 예상한다. 마지막으로 지금의 일자리 대체 담론이 갖는 문제점을 지적하고 현재의 사회기술적 실험을 느리지만 세밀하게 관찰하고 기록하는 태도가 무엇보다 필요하다고 주장한다.FALSEkc
Functional Expansion of the Skin Microbiome: A Pantothenate-Producing Rothia Strain Confers Anti-Inflammatory and Photoaging-Protective Effects
No full textThe functional landscape of the skin microbiome is largely defined by dominant genera such as Cutibacterium and Staphylococcus, whereas rare commensals remain poorly understood. In this study, we identified Rothia kristinae BF00107, a skin-resident strain with a complete pantothenate biosynthesis pathway, as a novel postbiotic candidate with distinct dermatological benefits. BF00107 fermentation filtrate suppressed pro-inflammatory cytokines (IL-1 beta and TNF-alpha) in keratinocytes and restored extracellular matrix homeostasis in UVB-irradiated fibroblasts by upregulating COL1A1 expression and reducing MMP-1 levels. Consistent with the observed phenotypes, transcriptomic profiling revealed a strain-specific signature characterized by downregulation and upregulation of the expression of inflammatory mediators and barrier- and ECM-associated genes, respectively. Comparative genomics and metabolite profiling confirmed BF00107 as a unique high-pantothenate producer. Supplementation with pantothenic acid reproduced the anti-inflammatory and barrier-supporting effects of the strain, confirming its role as a key effector metabolite. Furthermore, BF00107 passed standard safety assessments, including the Human Repeat Insult Patch Test (HRIPT), Ames, and irritation tests, supporting its suitability for human applications. These findings establish the pantothenate-producing R. kristinae BF00107 as the first functionally validated Rothia strain with anti-inflammatory and photoaging-protective properties. This study expands the functional scope of the skin microbiome and highlights rare commensals as valuable reservoirs for safe, strain-specific postbiotic development.TRUEsciescopu