GIST Scholar
Not a member yet
30271 research outputs found
Sort by
Selective formaldehyde condensation on phosphorus-rich copper catalyst to produce liquid C3+ chemicals in electrocatalytic CO2 reduction
No full textRecent advancements in the CO2 reduction reaction (CO2RR) target multicarbon chemical production and scalable electrode designs for industrial applications. Here we introduce a zero-gap cell utilizing humidified gas-phase CO2 and circulated alkaline media, achieving a Faradaic efficiency of 66.9% for C3+ products and a current density of −1,100 mA cm−2. In situ spectroscopic analyses revealed formaldehyde as a key intermediate formed on copper oxide/hydroxide interfaces derived from a phosphorus-rich copper catalyst. Unlike conventional pathways based on dimerization of CO intermediates, our study selectively produces liquid-phase multicarbon products because of autonomous local pH variations under a weak alkaline microenvironment, with allyl alcohol as the dominant C3+ product. The high selectivity and efficiency for liquid products provide a substantial advantage for storage and transport, highlighting the scalability and practical feasibility of our approach, which offers a potential economically viable solution for CO2 utilization. This development encourages the adoption of CO2RR technologies in iron–steel and petrochemical industries to mitigate greenhouse gas emissions. (Figure presented.) © The Author(s) 2025.TRUEsciescopu
Development of feature selection methods for cancer diagnosis and prognosis using omics data Euiyoung Oh School of Electrical Engineering and Computer Science Gwangju Institute of Science and Technology
No full textIn the field of cancer diagnosis and prognosis, identifying relevant feature subsets is essential to improve prediction model performance and determine biomarkers. In this dissertation, we propose two study approaches to identify significant genes associated with various types of cancer: (1) exploring the prognostic efficacy of both tumor and tumor-adjacent normal tissues, and (2) developing a deep neural network architecture for feature selection. In the first part of the dissertation, we present the prognostic efficacy of transcriptomic data from both tumor and adjacent normal tissues, utilizing The Cancer Genome Atlas (TCGA) dataset. By applying Cox regression models for prognostic analysis and machine learning models for survival prediction, the study demonstrates that for cancers such as kidney, liver, and head and neck, adjacent normal tissues exhibit a higher proportion of prognostic genes and outperform tumor tissues in survival prediction accuracy. Moreover, a distance correlation-based feature selection method was applied to external datasets for kidney and liver cancer, further confirming that genes selected from adjacent normal tissues consistently showed better prediction performance improvement than those from tumor tissues. These findings suggest that adjacent normal tissues may provide valuable insights into cancer prognosis, positioning them as potential targets for biomarker discovery. In the second part of the dissertation, we present a novel machine learning-based feature selection method called “Deep neural network with PaIrwise connected layers integrated with stochastic Gates” (DeepPIG) to address the challenge of selecting relevant features from complex omics data, particularly when feature signals are weak. Built upon the knockoff filter framework, DeepPIG is designed to enhance the detection power of relevant features without violating the false discovery rate (FDR) threshold. In comparison with baseline and recent models, such as Deep feature selection using Paired-Input Nonlinear Knockoffs (DeepPINK) and SHapley Additive exPlanations (SHAP), DeepPIG demonstrated superior detection power on synthetic datasets, particularly in cases where feature signals were subtle. Furthermore, in real-world applications, including cancer prognosis prediction and microbiome and single-cell data classification tasks, DeepPIG consistently outperformed traditional models in selecting relevant features and improving classification performance. The model’s robustness, especially when feature signals are weak, highlights its potential utility in a variety of high-dimensional biological data analyses. This dissertation highlights the potential of both tumor-adjacent normal tissues and the novel DeepPIG model as valuable tools in the field of cancer diagnosis and prog- nosis. These findings can enhance prognostic insights from high-dimensional biological data, improving model accuracy and supporting more precise biomarker discovery.DoctorAbstract i
감 사 의 글 iv
List of Contents vi
List of Tables viii
List of Figures ix
List of Algorithms x
1 Introduction 1
1.1 Introduction 1
1.2 Problem Statement 2
1.3 Proposed Approach 3
2 Background and Related Works 4
2.1 Tumor-adjacent normal tissues as cancer prognostic markers 4
2.2 Feature selection based on knockoff framework 5
3 Survival analysis using transcriptomic data revealed that tumor-adjacent
normal tissues harbor prognostic information on multiple cancer types 6
3.1 Materials and Methods 6
3.1.1 Study design 6
3.1.2 Datasets and preprocessing 8
3.1.3 Identification of differentially expressed genes and their expression ratio 9
3.1.4 Data screening via distance correlation 10
3.1.5 Survival prediction model and evaluation 10
3.1.6 Functional annotation 12
3.2 Results 12
3.2.1 Survival analysis with clinical data, gene expression data of tu-
mor and normal tissues, and expression ratio of DEGs 12
3.2.2 Prognostic values of selected features for kidney and liver cancer 16
3.2.3 Functional annotation of survival-related genes 19
3.3 Discussion 21
4 DeepPIG: deep neural network architecture with pairwise connected
layers and stochastic gates using knockoff frameworks for feature selection 24
4.1 Methods 24
4.1.1 Knockoff framework 24
4.1.2 Proposed Model 27
4.2 Simulation Studies 31
4.2.1 Synthetic data 31
4.2.2 Simulation results 32
4.3 Real Data Analysis 34
4.3.1 Transcriptomic Markers of Cancer Prognosis 34
4.3.2 Microbiome and single-cell datasets 38
4.4 Discussion 40
5 Supplementary Information 43
Summary 56
References 5
A Study on the Structural Reform of the National Pension and the Restructuring of the Old-age Income Security System
No full text본 논문은 세대 간 형평성을 높이면서 지속 가능한 노후소득보장제도 구축 방안을 제시하는 데 목적이 있다. 기초연금과 국민기초생활보장제도를 적용받는 저소득층 노인을 대상으로 최저소득보장을 도입해 조세를 통한 소득재분배 기능을 강화하고, 국민연금을 수지상등의 원칙에 기초한 30% 소득대체율의 제도로 전환하며, 소득대체율 10%의 프리미엄연금을 도입하고, 퇴직금제도를 퇴직연금으로 전환할 것을 제안한다. 현세대 노인의 빈곤은 최저소득보장을 통해 현세대 취업자가 책임지고, 미래세대 노인의 빈곤 예방은 수지상등이 준수되는 국민연금, 프리미엄연금 및 퇴직연금을 통해 이루어지기 때문에 초장수 사회를 책임져야 할 미래세대의 재정 부담이 큰 폭으로 감소하게 된다. 최저소득보장 도입에 필요한 비용은 기초연금을 유지하는 것과 비슷하며, 빈곤 감소 효과는 기초연금보다 훨씬 큰 것으로 분석되었다. 본 논문에서 제안한 방식으로 노후소득보장제도를 재구축하면 하위소득, 중위소득 및 상위소득 모든 계층의 노후소득이 증가하는 것으로 분석되었다. 이 구조개혁을 통해 우리 사회가 직면한 인구 위험, 저성장 위험 및 재정 위험을 세대 간에 분산시켜 세대 간 형평성을 유지할 수 있을 것으로 기대한다.FALSEkc
Harnessing Peptide Self-Assembly for Advanced Biomimetic Cryopreservatives
No full textPeptide self-assembly represents a powerful approach for developing functional biomaterials that mimic natural systems. We explore the application of this supramolecular strategy in creating biomimetic cryoprotectants inspired by ice-binding proteins. Among them, antifreeze peptides (AFPs) control ice crystal formation through precisely arranged binding domains, but their practical use is limited by extraction challenges and instability issues. Our research focuses on the rational design of short peptide sequences that self-assemble into nanostructures with enhanced ice recrystallization inhibition properties. These supramolecular assemblies create collective arrangements of ice-binding moieties that significantly amplify their interaction with ice crystals. By controlling supramolecular organization, we achieve superior cryoprotective properties compared to individual molecular units. We've also developed organic-inorganic nanohybrids, enabling precise control over ice-water interface interactions through size and shape modulation. The optical properties of these nanohybrids allow direct visualization of ice recrystallization inhibition, providing mechanistic insights. Our studies reveal that the collective behavior of functional groups within these self- assembled structures is crucial for enhanced performance. By matching spatial arrangements to ice crystal lattices, we optimize thermodynamic control of ice formation. These peptide assemblies demonstrate excellent efficacy in cellular cryopreservation with significantly reduced cytotoxicity compared to conventional agents like dimethyl sulfoxide. This versatile approach allows for customizable designs tailored to specific applications in cell preservation, biobanking, and pharmaceutical storage. By bridging natural AFP functions and synthetic systems through controlled self-assembly, we offer promising solutions to long-standing challenges in cryopreservation technologies
Analysis and fully memristor-based reservoir computing for temporal data classification
No full textReservoir computing (RC) offers a neuromorphic framework that is particularly effective for processing spatiotemporal signals. Known for its temporal processing prowess, RC significantly lowers training costs compared to conventional recurrent neural networks. A key component in its hardware deployment is the ability to generate dynamic reservoir states. Our research introduces a novel dual-memory RC system, integrating a short-term memory via a WOx-based memristor, capable of achieving 16 distinct states encoded over 4 bits, and a long-term memory component using a TiOx-based memristor within the readout layer. We thoroughly examine both memristor types and leverage the RC system to process temporal data sets. The performance of the proposed RC system is validated through two benchmark tasks: isolated spoken digit recognition and with only a fraction of complete samples forecasting the Mackey-Glass (MG) time series prediction. The system delivered an impressive 98.84% accuracy in speech digit recognition and sustained a low normalized root mean square error (NRMSE) of 0.036 in the time series prediction task, underscoring its capability. This study illuminates the adeptness of memristor-based RC systems in managing intricate temporal challenges, laying the groundwork for further innovations in neuromorphic computing. © 2024 Elsevier LtdFALSEsciescopu
RSCF: Relation-Semantics Consistent Filter for Entity Embedding of Knowledge Graph
No full textIn knowledge graph embedding, leveraging relation specific entity transformation has markedly enhanced performance. However, the consistency of embedding differences before and after transformation remains unaddressed, risking the loss of valuable inductive bias inherent in the embeddings. This inconsistency stems from two problems. First, transformation representations are specified for relations in a disconnected manner, allowing dissimilar transformations and corresponding entity embeddings for similar relations. Second, a generalized plug-in approach as a SFBR (Semantic Filter Based on Relations) disrupts this consistency through excessive concentration of entity embeddings under entity-based regularization, generating indistinguishable score distributions among relations. In this paper, we introduce a plug-in KGE method, Relation-Semantics Consistent Filter (RSCF). Its entity transformation has three features for enhancing semantic consistency: 1) shared affine transformation of relation embeddings across all relations, 2) rooted entity transformation that adds an entity embedding to its change represented by the transformed vector, and 3) normalization of the change to prevent scale reduction. To amplify the advantages of consistency that preserve semantics on embeddings, RSCF adds relation transformation and prediction modules for enhancing the semantics. In knowledge graph completion tasks with distance-based and tensor decomposition models, RSCF significantly outperforms state-of-the-art KGE methods, showing robustness across all relations and their frequencies. © 2025 Association for Computational Linguistics
Low-grade magnesium alloy scraps: An efficient and cost-effective reducing agent for the removal of Cr(VI) as a model contaminant
No full textThe lack of feasible technologies to recycle low-grade Mg alloy scrap (LGS) has led to its accumulation in landfills and manufacturing/recycling plants. To address this issue, we examined the feasibility of harnessing LGS as a powerful and cost-efficient reductant for hexavalent chromium (Cr(VI)) using AZ91 alloy scrap (Mg-AZS) as a representative LGS. After a preliminary wash with acetone to remove the lubricant oil retained in the scraps, Mg-AZS achieved 96.3 % Cr(VI) removal efficiency ([Cr(VI)]0 = 19.4 μM) within 48 h. The surface-area-normalized reactivity of Mg-AZS was at least comparable to, or even higher than, that of pure Mg(0), suggesting its potential as a cost-effective alternative. This enhanced reducing power was attributed to the composition of Mg-AZS, which inherently contains trace metal impurities (e.g., 0.24 wt%Fe and 0.12 wt%Cu), making it an intrinsic mixed-metal catalyst. Upon sonication before use, its reactivity increased 3-fold, achieving 97.5 % Cr(VI) removal within 6 h due to the effective removal of residual oil. Additionally, sonication alleviated the passivation of Mg-AZS caused by the surface precipitation of Cr(OH)3, enabling continuous reuse for at least 6 cycles. Finally, Mg-AZS achieved 100 % Cr(VI) removal in contaminated groundwater ([Mg(0)]0 = 0.25 g/L, [Cr(VI)0 = 500 μM]) within 40 min. These results demonstrate the potential of LGS for treating Cr(VI)-contaminated groundwater and reducing Mg waste. © 2025 Elsevier LtdFALSEsciescopu
Magnetic fluctuation and reversal by current-induced spin-orbit torques in heavy metal/ferromagnet heterostructure
No full textThe structure utilizing a heavy metal/ferromagnetic heterojunction induces spin polarization due to the spin Hall effect when current flows through the heavy metal layer. The resulting spin current transfers spin-orbit torque to the ferromagnet, leading to the phenomenon such as spin wave excitation or magnetic reversal. We confirmed the amplification of magnetic fluctuations in a nanostructure, Ta(1)/Pt(10)/Pt1-xCox(5)/Al2O3(5),co-deposited with Permalloy and cobalt through Brillouin light scattering (BLS) system. Although direct measurements of the auto-oscillation state could not be obtained due to resolution limitations, we indirectly observed that the critical current was reduced by 27.3% from the analysis of the BLS spectrum when Co was co-deposited compared to the case where only Permalloy was deposited. Current-induced field-free magnetic reversal using spin–orbit torque was conducted in heavy metals/ferromagnetic heterstructures. The ferromagnetic layer was designed as a [Co/Pt] multilayer structure to control perpendicular magnetic anisotropy of free layer instead of single ferromagnetic layer. Additionally, by exploiting the interlayer exchange interaction between a copper insertion layer and a ferromagnetic layer with in-plane magnetic anisotropy. the symmetry of the structure was broken, enabling magnetization reversal without the need for an external magnetic field. We confirm magnetic reversal electrically by anomalous Hall measurement in Hall bar pattern.This approach provides a simple method to optimize the efficiency of magnetization reversal in studies which utilize interlayer exchange interaction as symmetry-breaking component.DoctorAbstract.i
Contentsii
List of figures iii
Chapter 1. Review of the relevant literature.1
1. 1. Interaction in Magnetism1
1. 2. Magnetic Energy,,,3
1. 3. Spin wave8
1. 4. The ordinary, anomalous,and spin Hall effects.10
1. 5.Reference..,,,.12
Chapter 2 Effects of cobalt-permalloy cosputtering in a nanowire structure on magnetic fluctuation and auto-
oscillation.
2. 1. Introduction.19
2. 2. Experiment..20
2. 3. Results and discussion.20
2. 4. Conclusion..22
2. 5. Reference,,23
Chapter 3 Current-Induced Field-Free Switching of Co/Pt Multilayer via Modulation of Interlayer Exchange Coupling and
Magnetic Anisotropy
3. 1. Introduction 29
3. 2. Experiment31
3. 3. Results and discussion 31
3. 4. Conclusion34
3. 5. Reference,,34
CONCLUSION4
A R-RC Oscillator Based Temperature Sensor with Wide Temperature Range
No full text이 논문은 저항의 이차항의 영향를 보상하여 넓은 온도에서 동작 가능한 저항 기반 온도센서를 제시한다. 제안하는 swing-boosted R-RC 발진기는 두 가지 타입의 저항(poly 와 diffusion)을 사용한다. 두 가지 저항의 비율을 조절함으로써 저항의 이차항의 영향을 보상할 수 있다. 전원 민감도는 셀프-클락 차지 펌프를 갖는 레귤레이터를 사용함으로 완화된다. 더 높은 전력 효율성을 달성하기위해, 변환 시간을 4배 감소시키는 주파수 쿼드러플러가 사용되었다. 40nm CMOS 공정에서 제작된, 제안하는 센서는 152 μ s의 변환시간동안 32 μ W 를 소모하며 27.78mKLSB의 해상도를 달성한다. 이는 5pJ ∙ K2 의 해상도 FoM을 의미한다. 이 센서는 또한 2942μm2의 작은 면적에서 2점 교정과 3차항 피팅을 통해 넓은 온도 범위(- 60℃에서 160℃까지) 에서 -0.83/+1.31℃(3σ)의 부정확도를 달성한다. ©2025 Haseong Song ALL RIGHTS RESERVED MS/EC 20231138|This paper presents a resistor based temperature sensor capable of operating over a wide temperature range (from -60℃ to 160℃) by compensating second order influence of resistor. The proposed swing-boosted R-RC oscillator(R-RCOSC) employs two types of resistors (poly and diffusion). By adjusting the ratio of two resistors, second order influences of resistors are compensated. Supply sensitivity is suppressed by using regulator with self-clock charge pump. To achieve higher power efficiency, frequency quadrupler is used which reduce conversion time by a factor of four. Fabricated in a 40nm CMOS process, the proposed sensor achieves 27.78mKLSB while consuming 32μW in 152us conversion time. This corresponds resolution figure-of-merit(FoM) of 5pJ∙ K2. It also achieves inaccuracy of -0.83/+1.31℃(3σ) with 2-point calibration and 3rd order polynomial fitting, over wide temperature range (from -60℃ to 160℃) while occupying 2942μm2 small area. ©2025 Haseong Song ALL RIGHTS RESERVED MS/EC 20231138Master1. Introduction 1
2. Proposed Temperature Sensor 5
2.1 Sensing Mechanism 5
2.2 Proposed Swing-Boosted R-RC Oscillator 7
2.2.1 Conventional RC Oscillator 7
2.2.2 Proposed R-RC Oscillator 8
2.2.3 Method of second order coefficient compensation 11
2.3 Architecture of Regulator and Frequency Quadrupler 16
2.3.1 Regulator with Self-Clock Charge Pump 16
2.3.2 Frequency Quadrupler 18
3. Monte-Carlo Simulation Results 21
4. Conclusion 24
Reference 2
Prospective Associations of Serum Tumor Necrosis Factor-Alpha and Employment on Suicidal Behaviors Over 1 Year in Depressive Patients Receiving Psychopharmacotherapy
No full textObjective This study explored both the individual and combined effects of serum tumor necrosis factor-alpha (sTNF-alpha) levels and employment status on suicidal behavior (SB) in patients with depressive disorders undergoing pharmacologic treatment. Methods Baseline measurements of sTNF-alpha levels were taken, and employment status was determined. Over a 1-year period of stepwise pharmacotherapy, SB was monitored and categorized into increased suicidal severity and fatal/non-fatal suicide attempts. Logistic regression models adjusted for relevant covariates were used to analyze the individual and interactive associations between sTNF-alpha levels, employment status, and these two forms of SB. Results Unemployment was significantly associated with both forms of SB, whereas sTNF-alpha levels alone did not show a significant correlation. However, lower sTNF-alpha levels combined with employment were associated with the lowest incidence rates of both SB categories, demonstrating significant interactive effects after adjustment. Conclusion The study demonstrates that the prospective associations of sTNF-alpha levels for SB is enhanced when combined with employment status in patients receiving pharmacological treatment for depressive disorders. These findings suggest that integrating biological markers with socio-economic factors can improve the assessment and management of suicide risk. Psychiatry Investig 2025;22(7):748-756TRUEsciessciscopuskc