GIST Scholar
Not a member yet
30271 research outputs found
Sort by
Rotating Machinery Vibration Signal Augmentation Method Preserving Characteristic Frequency Regions
No full text회전 기계의 지속적인 진단을 위해 인공지능 방법이 적극적으로 활용되고 있으나, 산업 현장에서는 결함 데이터가 부족한 데이터 불균형 현상이 흔히 나타난다는 문제가 있다. 이를 해결하기 위해 데이터 증강 기법이 여럿 제안되어왔으나, 최소한의 개수 요구, 다양성 부족, 물리적 의미 훼손, 신호 특성 왜곡 등 여전히 한계가 있다. 본 연구에서는 특성 주파수 영역을 유지하는 회전체 진동 신호 증강 기법을 제안한다. 제안하는 방법은 a) 주파수 영역 이동 및 복원을 통한 비 특성 영역 신호 추출, b) 원래 신호와 비 특성 신호의 차를 이용한 특성 영역 신호의 추출, c) 서로 다른 비 특성 영역 신호와 특성 영역 신호의 합을 통한 새로운 데이터 생성의 과정으로 증강을 수행한다. 해당 방법으로 획득한 데이터를 특성인자 분포 비교, CNN 기반 진단 모델을 통해 다른 증강 기법과 비교하여 평가하고 검증하였으며, 기존 방법에서 제시된 한계점을 해결하여 실제 데이터와 유사한 특성을 가지면서도 다양성이 부여된 고품질 증강데이터를 만들 수 있음을 확인하였다.MasterAbstract i
Contents iii
List of Figures iv
List of Tables vi
Chapter 1. Introduction 7
Chapter 2. Theoretical Background and Literature Review 10
2.1. Requirements for Vibration Signal Augmentation Method 10
2.2. Previous Vibration Signal Augmentation Methods 12
Chapter 3. Proposed Vibration Signal Augmentation Method 14
3.1. Signal Decomposition Method to Target Signal and Residual Signal 14
3.2. Overall Process of Augmentation Method 17
3.3. Augmented Data Evaluation Methods 19
3.4. Diagnosis Framework for Class Imbalanced dataset Situation 22
Chapter 4. Case Studies 26
4.1. Case Study 1: Simulation Data 26
4.2. Case Study 2: SDDO Testbed Data 36
4.3. Case Study 3: KONA EV Drivetrain Data 40
Chapter 5. Conclusions and Future Works 44
Appendix 46
Appendix A. Python Code of Proposed Augmentation Method 46
Appendix B. Hyper parameter study 50
References 52
Curriculum Vitae 5
Multiscale topology optimization of electropermanent magnet composites in magnetic actuators
No full textElectropermanent magnet (EPM) composites are engineered structures that integrate permanent magnets (PMs) and iron at the microscale. This work aims to demonstrate that a graded EPM microstructure, designed through multiscale topology optimization, enhances the magnetic force of actuators beyond that of traditional single-scale designs. Specifically, a homogenization-based multiscale topology optimization approach is employed in this work, consisting of three steps: (1) constructing a surrogate model of effective material properties, (2) determining the optimal distribution of macroscopic and microstructure design variables, and (3) reconstructing the optimized EPM microstructure at the macroscale. The effectiveness of this approach is evaluated through three numerical examples. The first example provides the design results at various iron-to-PM volume ratios. The second example quantitatively compares the magnetic forces of multiscale and single-scale designs, demonstrating the superiority of the multiscale approach. The third example investigates the effect of EPM unit cell rotation and PM magnetization direction change. This study confirms the potential of multiscale topology optimization in addressing magnetic field problems for electromechanical systems.FALSEsciescopu
Preferred Jacobian Differentiation and Direct Collocation Methods for an Efficient and Accurate Walker Gait Optimization
No full textDevices based on walker robotics research sometimes require a reference trajectory for the control systems of the devices to track. One valuable class of methods used to find an optimal reference trajectory is direct collocation, but even after selecting a method like direct collocation, several optimization design decisions remain. In order to determine the most desirable optimization settings, 600 optimizations were performed for the trajectory of a two degree-of-freedom (DOF) compass gait walker, and 200 optimizations were performed for a five-DOF link walker. These runs evaluated various combinations of optimization settings, including: numerical vs. symbolic vs. automatic differentiation; trapezoidal vs. Hermite Simpson collocation; numerical vs. symbolic calculation of joint accelerations; and inclusion or exclusion of joint accelerations in the decision variables. The different generated gaits were then compared in terms of computational efficiency and accuracy. The results showed that including joint accelerations as decision variables eliminated alternative gaits but increased computational complexity and variability. Symbolic acceleration evaluation was preferable when automatic differentiation was excluded. Automatic differentiation was shown to be significantly faster than the other two differentiation methods for both walking models. In addition, Hermite-Simpson collocation, although slower than trapezoidal, was the more accurate of the two approaches. These results can be applied to the derivation of optimal reference joint trajectories in future robotics applications.FALSEsciescopuskc
Reproductive functions of venerose, a sexually transmitted sugar in the fruit fly Drosophila melanogaster
No full textSeminal fluid is more than just carrier for sperm. Its complex composition, consisting of ions, peptides, proteins, lipids, and carbohydrates, actively communicates with the female reproductive system. However, the significance of non-protein substances such as present in seminal fluid remains largely underexplored. In this thesis, I utilized Drosophila melanogaster to investigate the influence of a seminal sugar on female reproduction. I found that a significant amount of phospho-galactoside named ‘venerose,’ present in the male seminal fluid, is incorporated into female ovaries and enters the haemolymph. The findings revealed that venerose acts as a signaling molecule for promoting the mating-induced germline stem cell (GSC) proliferation and enhancing sperm storage, particularly in undernourished females. Circulating venerose directly activates Dh44-PI neurons, nutrient-sensing cells in the brain. The secreted Dh44 stimulates GSC proliferation by activating its receptor in terminal filament cells and enhancing Decapentaplegic (Dpp) signaling. Additionally, the evidence indicates that undernourished females exhibit the increased Dh44 release in response to venerose. This increased Dh44 secretion delays the removal or expulsion of the ejaculate (sperm ejection), thereby increasing the sperm uptake and venerose absorption. This study establishes a framework for understanding the function of nutrient-like substances abundant in seminal fluid and provides mechanical insights into how females assess courtship feeding for sexual selection.DoctorAbstract
List of contents
List of figures and tables
Chapter 1. General introduction
1.1 The role of seminal plasma in enhancing male fertility
1.2 Complex roles of seminal fluid in Drosophila melanogaster reproduction
1.3 Germline stem cell proliferation in Drosophila melanogaster female
1.4 Sexual selection and courtship feeding
Chapter 2. Identification of venerose, a sexually transmitted sugar in the fruit fly and its reproductive functions
2.1 Introduction
2.2 Results
2.2.1. Venerose, a phospho-galactoside abundant in seminal fluid
2.2.2. Glucosyltransferase is required for venerose production
2.2.3. Females absorb venerose from the male ejaculate
2.2.4. Venerose is essential for mating-induced GSC proliferation
2.2.5. Venerose stimulates GSC proliferation via a brain factor
2.2.6. Dh44 acts on the ovaries to trigger GSC proliferation
2.2.7. Venerose stimulates GSC proliferation via Dh44-PI neurons
2.2.8. Venerose stimulates Dh44-PI neurons to secrete Dh 44
2.2.9. Venerose stimulates Dh44-PI neurons via Tret1- 1
2.2.10. Energy-deprived females absorb more venerose
2.2.11. Energy-dependent effects of venerose on EHP and sperm storage
2.2.12. Energy state determines Dh44 pool size
2.2.13. Females assess the energy status of male partners via venerose
2.2.14. Other functions of venerose in female systems
2.3. Discussion
2.3.1. Venerose is a signaling molecule that activates the Dh44 system during mating
2.3.2. Dh44 regulates GSC proliferation and EHP expansion via distinct GPCR pathways
2.3.3. Venerose, a non-protein seminal substance that promotes GSC proliferation
2.3.4. Venerose transfer is a form of courtship feeding signaling male quality
2.3.5. Nutritional stress enhances acceptance of courtship feeding
2.3.6. The Dh44 system couples oogenesis with energy state
2.3.7. A potential implication in mammalian systems
2.3.8. Limitations of this study
2.4. Material and Methods
Chapter 3. General Conclusion
Chapter 4. References
List of figures and tables
Figures
Figure 1. Interplay of nutritional and seminal signals regulating germline stem cell proliferation and reproductive responses in Drosophila melanogaster
Figure 2. Identification of venerose, a sugar-like substance rich in male Drosophila ejaculate
Figure 3. Structural analysis of native venerose
Figure 4. Venerose production requires the 1,2-Diacyl-glycerol (DAG) biosynthesis genes or lipase genes in the MAG
Figure 5. MAG-specific knockdown of the glycosyltransferase UGT305A1 inhibits venerose production
Figure 6. UGT305A1 knockdown had limited effects on MAG morphology or SP content
Figure 7. MAG-specific knockdown of the glycosyltransferase UGT305A1 reduces phosphorus accumulation in the ovaries of females mated with knockdown males
Figure 8. Venerose, enhancing egg-laying activity, has a limited impact on sperm fertility or offspring survival rate
Figure 9. Females mated with UGT305A1 knockdown males show defects in GSC proliferation, but this effect can be restored by venerose injection
Figure 10. Venerose stimulates GSC proliferation by boosting BMP signaling activity to GSCs
Figure 11. Venerose stimulates GSC proliferation via the brain
Figure 12. Dh44-R2 and Bab1-Gal4 expression do not overlap in the ovary
Figure 13. The neuropeptide Dh44, secreted by brain Dh44-PI neurons, and its receptor Dh44-R2, expressed in GSC niche cells, are essential for venerose-induced GSC proliferation
Figure 14. Dh44-R2 knockdown females show no increase in GSC proliferation upon mating
Figure 15. Venerose stimulates Ca2+ transients and Dh44 secretion in Dh44-PI neurons via the sugar transporter Tret1- 1
Figure 16. Females with Tret1-1 knockdown in Dh44-PI neurons exhibit do not differentiate between nutritive D-glucose and non-nutritive L-glucose
Figure 17. Nutritional stress in females extends EHP duration and enhances venerose absorption
Figure 18. Starvation decreases the neural activity of Dh44-PI neurons
Figure 19. Nutritional stress in females extends EHP duration by increasing the Dh44 pool available for venerose-induced secretion
Figure 20. Under nutritional stress conditions, females discern the energy states of their mates by extending EHP, absorbing more venerose, and storing a greater amount of sperm from well-nourished males compared to undernourished males
Figure 21 Females mated with venerose-depleted males partially decreases mating-induced vitellogenesis, which is fully rescued by venerose injection
Figure 22. Venerose is not utilized as an oral energy source in females
Figure 23. Venerose can alters gut gene expression, and defecation activity
Figure 24. Starved females mated with well-nourished males exhibit stronger GSC proliferation upon mating than those mated with starved males. Oamb-deficient females show GSC proliferation in response to synthetic venerose injectio
Enhancing multi-task in vivo toxicity prediction via integrated knowledge transfer of chemical knowledge and in vitro toxicity information
No full textThe evaluation of potential drug toxicity is a crucial step in early drug development. in vivo toxicity assessment represents a key challenge that must be addressed before advancing to clinical trials. However, traditional in vivo experiments primarily rely on animal models, raising concerns regarding cost, time efficiency, and ethical considerations. To address these challenges, various computational approaches have been developed to support in vivo toxicity evaluations, though these methods often demonstrate limited generalizability due to data scarcity. In this study, we propose MT-Tox, a knowledge transfer-based multi-task learning model specifically designed for in vivo toxicity prediction that overcomes data scarcity. Our model implements a sequential knowledge transfer strategy across three stages: general chemical knowledge pretraining, in vitro toxicological auxiliary training, and in vivo toxicity fine-tuning. This hierarchical approach significantly improves model performance by systematically leveraging information from both chemical structure and toxicity data sources. MT-Tox outperforms baseline models across three in vivo toxicity endpoints: carcinogenicity, drug-induced liver injury (DILI), and genotoxicity. Through ablation studies and attention analyses, we demonstrate that each knowledge transfer technique makes meaningful contributions to the prediction process. Finally, we demonstrate the real-world application of our model as a prediction tool for early-stage drug discovery through comprehensive DrugBank database screening. Scientific contribution: We propose a knowledge transfer framework that integrates chemical and in vitro toxicological information to enhance in vivo toxicity prediction in low-data regimes. Our model provides dual-level interpretability across chemical and biological domains through attention mechanism. Moreover, we demonstrate our model’s applicability by screening the DrugBank database, simulating practical toxicity screening scenarios in drug development. © The Author(s) 2025.TRUEsciescopu
Amine-Modified SPEEK Membranes via Interfacial Polymerization for Li+/Mg2+ Separation in Electrically Driven Systems
No full textLithium is a critical resource essential for energy storage, yet its selective extraction from salt-lake brines remains challenging due to the low Li+ concentration and the high Mg2+/Li+ ratio. This study investigates the potential of sulfonated poly(ether ether ketone) (SPEEK) as a substrate material for lithium ion (Li+) separation in electrically driven systems. SPEEK membranes were prepared with varying sulfonation reaction times (12, 24, and 36 h) to identify the optimal conditions by evaluating their mechanical properties, ionic flux, and Li+ selectivity. The SPEEK-12 membrane exhibited superior performance and was utilized as a substrate for interfacial polymerization with amine monomers, including polyethylenimine (PEI), piperazine (PIP), and m-phenylenediamine (MPD), which were cross-linked with 1,3,5-benzenetricarbonyl trichloride (TMC). This process formed positively charged thin-film composite layers, enhancing Mg2+ rejection and Li+/Mg2+ selectivity. The modified membranes were extensively characterized using ATR FT-IR, XPS, SEM, and AFM to confirm chemical and morphological changes. PIP-TMC-SPEEK exhibited the highest Li+/Mg2+ selectivity (434.01) and Li+ flux (48.75 mmol m–2 h–1), exceeding the performance of other membranes under simulated natural brine conditions. These findings demonstrate that SPEEK membranes modified via interfacial polymerization are promising candidates for efficient and selective Li+ extraction in electrically driven systems.FALSEsciescopu
Multi-task learning-based temporal pattern matching network for guitar tablature transcription
No full textGuitar tablature transcription poses unique challenges in automatic music transcription, as it requires capturing both pitch and string usage on a multi-string instrument with various expressive techniques. While guitar tablature is widely used by guitarists in the music field, neural architecture modeling for this task remains underexplored, particularly in accurately mapping pitches to their respective strings. In this work, we propose a multi-task learning-based temporal pattern-matching network (TPMNet) that effectively captures temporal information from guitar recordings, improving the alignment of predicted results. The key contribution of this work is the advancement of neural network architecture, leading to notable improvements in prediction performance for guitar tablature transcription. Additionally, we explored the optimal pooling layer selection method tailored to different tasks, addressing a long-confusing problem in the field. TPMNet’s efficacy was validated through experiments on the GuitarSet dataset, and its generalizability was confirmed via cross-evaluation with the EGDB dataset. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.FALSEscopu
Pharmacological Reprogramming of the Tumor Microenvironment Through Targeted Inhibition of Immune and Stromal Interactions
No full text종양 미세 환경(TME)은 암세포, 면역 세포 및 기질 세포 간의 복잡한 상호작용을 통해 종양 성장, 면역 회피 및 전이에 중추적인 조절 효과를 발휘합니다. 특히, IL-6는 암 연관 섬유모세포(CAF)에서 분비되며, 이 사이토카인이 종양 연관 대식세포(TAM)를 면역 억제성 M2 유형으로 분극화하여 불리한 면역 환경을 조성한다는 관찰이 주목할 만한 연구 결과입니다. 이 연구에서는 IL-6를 억제할 수 있는 메커니즘을 확인하기 위해 FDA 승인 약물 라이브러리를 사용하여 약물 스크리닝을 수행했습니다. 이 스크리닝을 통해 화합물 #4065를 확인하였고, 이 화합물은 IL-6 분비를 강력하게 억제할 뿐만 아니라 TAM을 M1 유형으로 효과적으로 재프로그램하여 NK 세포의 면역 활성을 회복시키는 능력을 보여주었습니다. 이러한 연구 결과는 이 화합물이 마우스 종양 모델에서 종양 억제 효과를 나타낸다는 것을 시사합니다.
반대로, M2 유형 TAM이 분비하는 케모카인인 CCL22는 CAF 표현형 유도의 새로운 매개체로 밝혀졌습니다. CCL22는 α-SMA 및 FAP 발현을 유도하는 동시에 PDGFRβ 발현을 억제하는 것으로 나타났습니다. 또한, CCR4 수용체를 통해 섬유모세포가 CAF로 전환하는 것을 촉진하는 것으로 관찰되었습니다. 이러한 효과는 CCR4 특이적 억제제인 AZD2098에 의해 억제되는 것으로 밝혀졌으며, 이는 CCL22-CCR4 축이 TAM-CAF 상호 작용에서 중추적인 역할을 한다는 것을 시사합니다. 이 발견은 TME를 조절할 수 있는 새로운 치료 표적을 확립했습니다.
노화에 따른 면역 미세환경의 변화도 중요한 역할을 합니다. TCGA-KIRC 데이터를 분석한 결과, 노화된 암환자의 정상 조직에서 이중 특이적 포스파타제인 DUSP22의 발현이 지속적으로 증가하며, 이는 종양 진행과 직접적인 관련이 있는 것으로 밝혀졌습니다. 실험적 증거에 따르면 노화된 마우스에서 추출한 대식세포에서 DUSP22 발현이 증가하여 신세포암종(RCC)의 침습성이 크게 향상되는 것으로 나타났습니다. 또한 DUSP22는 종양 세포의 침입과 전이에 관여하는 것으로 밝혀져 종양 발생에 기여할 수 있는 가능성을 시사했습니다. siRNA로 DUSP22를 녹다운하거나 BML-260과 같은 DUSP22 억제제로 치료하면 침습성이 현저히 감소하는 것으로 나타났습니다. 이러한 연구 결과는 DUSP22가 노인 환자의 종양 악성화에 기여하는 중요한 요인으로 밝혀졌습니다.
요약하면, 본 연구는 CAF-IL6-TAM, CCL22-CCR4-CAF, DUSP22-TAM의 세 축이 종양 미세환경의 면역 억제 및 종양 촉진 특성을 조화롭게 조절하며, 이를 동시에 표적으로 하는 다중 접근법이 면역 기반 항암 치료 전략의 새로운 방향성을 제시할 수 있음을 보여줍니다.|The tumor microenvironment (TME) exerts a pivotal regulatory influence on tumor growth, immune evasion, and metastasis through intricate interactions among cancer cells, immune cells, and stromal cells. Of particular relevance is the observation that interleukin-6 (IL-6) is secreted in substantial amounts by cancer-associated fibroblasts (CAFs), and this cytokine polarizes tumor-associated macrophages (TAMs) into the immunosuppressive M2 type, resulting in an unfavorable immune environment. In this study, a drug discovery screen was conducted using a library of FDA-approved drugs to identify mechanisms that could inhibit IL-6. This screen resulted in the identification of compound #4065. The compound demonstrated the ability to not only potently inhibit IL-6 secretion, but also to effectively reprogram TAMs to the M1 type, thereby restoring the immune activity of NK cells. These findings suggest that the compound exhibits tumor-suppressing effects in a mouse tumor model.
Conversely, CCL22, a chemokine secreted by M2-type TAMs, has been identified as a novel mediator of CAF phenotype induction. CCL22 has been shown to induce α-SMA and FAP expression, while concurrently inhibiting PDGFRβ expression. Furthermore, it has been observed to promote the transformation of fibroblasts into CAFs via the CCR4 receptor. These effects were found to be inhibited by the CCR4-specific inhibitor AZD2098, suggesting that the CCL22-CCR4 axis plays a pivotal role in the TAM-CAF interaction. This finding establishes a novel therapeutic target for modulating the TME.
Changes in the immune microenvironment with aging also play an important role. An analysis of TCGA-KIRC data revealed a consistent increase in the expression of DUSP22, a dual-specific phosphatase, in aged normal tissues, which was directly associated with tumor progression. Experimental evidence has shown elevated DUSP22 expression in macrophages derived from aged mice, leading to a significant enhancement in the invasiveness of renal cell carcinoma (RCC). In addition, DUSP22 was shown to be involved in the invasion and metastasis of tumor cells, suggesting its potential to contribute to tumor development. Knockdown of DUSP22 with siRNAs or treatment with a DUSP22 inhibitor, such as BML-260, led to a significant reduction in invasiveness. These findings identify DUSP22 as a critical factor contributing to tumor malignancy in elderly patients.
In summary, the present study demonstrates that the three axes of IL-6-CAF-TAM, CCL22-CCR4-CAF, and DUSP22-TAM coordinately regulate the immunosuppressive and tumor-promoting properties of the tumor microenvironment, indicating that a multiplex approach to target them simultaneously may represent a novel direction for immune-based anticancer therapeutic strategies.DoctorPART 1. Immune Reprogramming of the Tumor Microenvironment through IL-6 Inhibition Using a Drug Repurposing 1
1. INTRODUCTION 2
2. MATERIALS AND METHODS 5
2.1 Reagents 5
2.2 Cell culture 5
2.3 Collection of conditioned media 5
2.4 Real-time quantitative polymerase chain reaction (PCR) 6
2.5 Enzyme-linked immunosorbent assay (ELISA) 6
2.6 MTT assay 6
2.7 Macrophage differentiation and polarization for THP-1 7
2.8 Transwell indirect co-culture 7
2.9 Animal studies 7
2.10 Mouse CAF purification and culture 8
2.11 Immunocytochemistry 8
2.12 Syngeneic colon cancer mouse model 8
2.13Magnetic-Activated Cell Sorting (MACS) 9
2.14 Flow cytometry 9
2.15 Statistical analysis 10
3. RESULTS 11
3.1 Repurposing of Antihypertensive Drug #4065 as a Potent IL-6 Inhibitor in CAF 11
3.2 Dose-Dependent Reprogramming of TAM by Compound #4065 in a CAF–Macrophage Co-culture System 11
3.3 Restoration of NK Cell Activity by Compound #4065 in a CAF–Macrophage–NK Cell Triple Co-culture Model 12
3.4 In Vivo Antitumor Efficacy of Compound #4065 in a CAF-Enriched Syngeneic Mouse Colon
Cancer Model 13
3.5 Flow Cytometric Analysis Reveals Broad Immunosuppressive Effects of Compound #4065 on TAMs and NK Cells in a CAF-Enriched Tumor Microenvironment 14
4. DISCUSSION 16
PART 2. Identification of CCL22 as a Key TAM-Derived Modulator of CAF Activation in the Tum
or Microenvironment 33
1. INTRODUCTIOIN 34
2. MATERIALS AND METHODS 36
2.1 Reagents 36
2.2 Cell culture 36
2.3 Macrophage differentiation and polarization for THP-1 36
2.4 Real-time quantitative polymerase chain reaction (PCR) 37
2.5 Collection of conditioned media 37
2.6 Cytokine antibody array 38
2.7 Immunocytochemistry 38
2.8 Animal studies 38
2.9 Xenograft colon cancer mouse model 38
2.10 Western blotting 39
2.11 Statistical analysis 39
3. RESULTS 40
3.1 PMA-Induced Differentiation and Cytokine-Mediated M1/M2 Polarization of THP-1 Cells 40
3.2 Opposing Roles of M1 and M2 Macrophages in Modulating Fibroblast Phenotype and Activation 40
3.3 Comparative Analysis of Cytokine Secretion Patterns in M1, M2, and TAM Phenotypes 41
3.4 Functional Role of CCL22 in Initiating and Shaping CAF Phenotypes in Fibroblasts 42
3.5 Inhibition of CCL22-Induced CAF Activation via Targeting the CCR4 Signaling Axis in Fibroblasts 43
3.6 In Vivo Validation of M2 TAM-Induced Tumor Progression and CAF Activation via the CCL22–CCR4 Axis 44
4. DISCUSSION 46
PART 3. Targeting DUSP22 in aged macrophages suppresses tumor invasion and metastasis 63
1. INTRODUCTIOIN 64
2. MATERIALS AND METHODS 66
2.1 Reagents 66
2.2 Cell culture 66
2.3 Collection of conditioned media 66
2.4 siRNA-mediated gene knockdown 66
2.5 Real-time quantitative polymerase chain reaction (PCR) 67
2.6 Transwell invasion assay 67
2.7 Immunocytochemistry 67
2.8 The enzyme-linked immunosorbent assay (ELISA) 68
2.9 Bone-marrow Derived Monocyte (BMDM) isolation 68
2.10 Zebrafish–Mouse Cancer Xenograft Model 69
2.11 Statistical analysis 69
3. RESULTS 70
3.1 Establishment of a Macrophage-Based Model for Validating Aging-Related Genes 70
3.2 Age-Related Phenotypic Shifts and Gene Expression Changes in Tumor-Associated Macrophages 70
3.3 Pro-Invasive Role of Aged Macrophages and Functional Characterization of DUSP22 in Tumor Cell Invasion 71
3.4 Pharmacological Targeting of DUSP22 Reduces Tumor Invasion and Metastasis in Cellular and Zebrafish Models 72
4. DISCUSSION 73
REFERENCES 90
ABSTRACT IN KOREAN 98
ACKNOWLEDGEMENT 9
Comparative analysis of directional cue perception between vibrotactile and electrotactile using funneling illusion mapping methods
No full textTactile illusion is a promising concept for navigation, sensory augmentation for assistance, virtual reality (VR), and mobile haptic interaction applications enabling efficient feedback with a limited number of actuators. Through the funneling illusion, directional cues can be generated using a tactor array-based system, transmitting cues even at locations where actuators are not physically mounted. This technique has been applied to various research objectives according to modality, mapping methods, body part, and other factors. Accordingly, experimental studies analyzing the perceptual characteristics of the funneling illusion mapping methods across different modalities are necessary. We examined the effectiveness of funneling illusion mapping methods using vibrotactile and electrotactile modalities on the trunk and shank. The study involved two types of experiments, circle tracking (continuous cue) and directional accuracy (discrete cue) experiments with 20 healthy participants. The experimental results showed that for the electrotactile modality, Power law-based mapping provided the most accurate and continuous cue on the shank for both cue types. Square and Power law-based mappings provided the most accurate and continuous cue on the trunk during the circle tracking experiment. For the vibrotactile modality, Linear mapping provided the most accurate and continuous cues during the circle tracking experiment. These results demonstrate that mapping method performance varies across experimental conditions. Our findings can guide the selection of experimental parameters when utilizing a tactor array-based system for the trunk and shank. © 2025 Elsevier B.V., All rights reserved.TRUEsciescopu
Analysis of a Memcapacitor-Based Online Learning Neural Network Accelerator Framework
No full textData-intensive computing tasks, such as training neural networks, are fundamental to artificial intelligence applications but often demand substantial energy resources. This study presents a novel complementary metal-oxide-semiconductor (CMOS)-based memcapacitor framework designed to address these challenges by enabling efficient and robust neuromorphic computing. Utilizing memcapacitor devices, a crossbar array that performs parallel vector-matrix multiplication operations, validated through cadence simulations and implemented in python for scalable accelerator design, is developed. The framework demonstrates outstanding performance across classification tasks, achieving 98.4% accuracy in digit recognition and 85.9% in object recognition. A key aspect of this research is its focus on real-world fabrication nonidealities, including up to 30% device parameter variations, ensuring robustness and reliability under practical deployment conditions. The results emphasize the effectiveness of capacitance-based systems in handling classification tasks while demonstrating resilience to fabrication-induced variations. This work establishes a foundation for scalable, energy-efficient, and robust memcapacitor-based neural networks, advancing the potential for intelligent systems in artificial intelligence-driven applications and paving the way for future innovations in neuromorphic computing. © 2025 The Author(s). Advanced Intelligent Systems published by Wiley-VCH GmbH.TRUEsciescopu