Search

Your search keyword '"Ahn, Jaehyun"' showing total 144 results
Start Over Author "Ahn, Jaehyun"
144 results on '"Ahn, Jaehyun"'

3. Understanding normalization in contrastive representation learning and out-of-distribution detection

Author

Le-Gia, Tai and Ahn, Jaehyun

Subjects
Computer Science - Computer Vision and Pattern Recognition, Statistics - Machine Learning
Abstract

Contrastive representation learning has emerged as an outstanding approach for anomaly detection. In this work, we explore the $\ell_2$-norm of contrastive features and its applications in out-of-distribution detection. We propose a simple method based on contrastive learning, which incorporates out-of-distribution data by discriminating against normal samples in the contrastive layer space. Our approach can be applied flexibly as an outlier exposure (OE) approach, where the out-of-distribution data is a huge collective of random images, or as a fully self-supervised learning approach, where the out-of-distribution data is self-generated by applying distribution-shifting transformations. The ability to incorporate additional out-of-distribution samples enables a feasible solution for datasets where AD methods based on contrastive learning generally underperform, such as aerial images or microscopy images. Furthermore, the high-quality features learned through contrastive learning consistently enhance performance in OE scenarios, even when the available out-of-distribution dataset is not diverse enough. Our extensive experiments demonstrate the superiority of our proposed method under various scenarios, including unimodal and multimodal settings, with various image datasets.

Published
2023

6. Case Studies of Female-Headed Farms and Households in Liberia: A Comparative Analysis of Grand Bassa, Lofa, and Nimba Counties

Author

Ahn, Jaehyun, Briers, Gary, Kibriya, Shahriar, and Price, Edwin

Abstract

Purpose: This study explores female-headed households in Grand Bassa, Lofa and Nimba counties to discern Liberia's smallholding, subsistence agriculture. Amid environmental and communal dynamics, addressing factors causing challenges of farming is imperative. Methodology: Using explanatory sequential methods this study collects, explains, and organizes farming and household situations qualitatively through field notes, observations, and extensive communications with 44 female farmers. Subsequently and quantitatively, Chi-Square Automatic Interaction Detection (CHAID) verifies aforementioned information, accompanied by 112 observations. Findings: Stories encapsulate female farmers in Grand Bassa, Lofa, and Nimba counties. They provide evidence of marginalized, progressive, and emerging farming in order. Children (n > 3) and inadequate support of Kuu (informal labor) and of community make food stability, availability and access insecure. Insufficient crop revenue further exacerbates the issues. Moderate- and mild food-insecure households need not worry about the problems above. However, their earnings notwithstanding, new challenges appear in credit practices and land conflict. Practical Implications: Extension agents should provide timely services for each community, with special attention to women-headed households. Findings imply ex-post evaluations to ex-ante extension services should be cyclical. Theoretical Implications: A gender-inclusive, gender-sensitive framework is vital to enhance food security and community resilience, transform subsistence agriculture into agribusiness and achieve fairness, gender equity and social justice for agrarian Liberians -- especially women and children. Originality/Value: This study captures distinctive phases of farming and family life -- yielding qualitative and subsequent quantitative validation of predictors of food insecurity in rural, women-headed farming households/operations. It should guide us to tailor ex-ante extension services.

Published
2020
Full Text
View/download PDF

10. Sustaining Global Food Systems with Youth Digital Livestock Production Curricula Interventions and Adoption to Professionally Develop Agents of Change

Author

Strong, Robert, primary, Sprayberry, Sarah, additional, Dooley, Kim, additional, Ahn, Jaehyun, additional, Richards, Jennifer, additional, Kinsella, Jim, additional, Lee, Chin-Ling, additional, Ray, Nicole, additional, Cardey, Sarah, additional, Benson, Carmen, additional, and Ettekal, Andrea, additional

Published
2023
Full Text
View/download PDF

14. A Novel Lateral Dynamics Control Strategy of In-Wheel Motor Vehicle to Improve Agility and Straight-Line Driving Stability

Author

Lee, Sangyeop, Kim, Keonchang, Moon, Seongjoon, Kim, Byongsung, Ahn, Jaehyun, Hwang, Junha, Kim, Donghyun, Woo, Seunghoon, and Yoo, Jinwoo

Abstract

In-wheel motor (IWM) vehicles offer a more sporty driving experience but suffer from left-right torque differences that impair straight-line driving stability. This paper proposes an integrated yaw-rate control strategy for IWM vehicle based on adaptive sliding mode control (ASMC) to enhance both agility in cornering and stability in straight-line driving situations. Given that 1) internal and external driving conditions such as driving maneuvers, road surfaces, and unwanted external disturbances affect vehicle yaw dynamics and 2) the tire cornering stiffness and understeer gradient are decisive parameters in yaw dynamics, the proposed strategy adapts online to these two parameters. Moreover, the proposed strategy adopts an adaptive update rate during the adaptation process to ensure robust disturbance rejection performance under various driving conditions. All control laws are defined solely based on measurable information in mass-production vehicles, without any knowledge of road-tire conditions or uncertainty bounds. In experimental tests, the proposed strategy shows improved control precision, accuracy, and robustness compared to PD controller and non-adaptive SMC. As a result, the proposed adaptation strategy improves the maneuverability of IWM vehicles in both dynamic cornering and straight-line driving situations.

Published
2024
Full Text
View/download PDF

20. Integrated Single Inductor Converter for a DC Grid System Connected With Battery and Load

Author

Jae-Kuk Kim, Seung-Hoon Lee, Ji-Yeon Kim, Bom-Seok Lee, and Ahn Jaehyun

Subjects
Computer science, business.industry, Mode (statistics), Electrical engineering, Battery (vacuum tube), High power density, Converters, Inductor, Magnetomechanical effects, Control and Systems Engineering, Grid system, Electrical and Electronic Engineering, business, Power density
Abstract

Among the various dc grid system concepts connected to battery and load, the parallel-connected structure is widely adopted due to its simplicity. However, it has two inductors, which operate simultaneously when using a battery, so the power density and efficiency can be degraded. To alleviate these problems, in this article, we propose a new single inductor converter. By integrating the two converters and using the battery switching cell, the proposed converter can reduce the number of inductors required and employ lower voltage-rated switches. Therefore, the proposed converter can achieve high power density and efficiency. A 500-W prototype converter is implemented, and each mode is examined under various operating conditions to confirm the validity.

Published
2021
Full Text
View/download PDF

21. Two-Mode Low-Voltage DC/DC Converter With High and Wide Input Voltage Range

Author

Jae-Kuk Kim, Bom-Seok Lee, Dae-Hun Kwon, Seung-Hoon Lee, and Ahn Jaehyun

Subjects
Forward converter, Materials science, business.industry, Electrical engineering, Inductor, law.invention, Capacitor, Main battery, Control and Systems Engineering, law, Electrical and Electronic Engineering, Transformer, business, Low voltage, Voltage, DC bias
Abstract

It has been proposed that the main battery voltage of electric vehicles should be increased to increase the driving distance and decrease the charging time. Thus, a low-voltage dc/dc converter (LDC) with high and wide input voltage range is required for electronic loads. This article proposes a two-mode LDC and the entire input range is divided into two low and high input voltage ranges. The converter has two operating modes corresponding to these input voltage ranges. In the low input voltage range, it operates as a three-switch double-ended active clamp forward converter. In the high input voltage range, it operates as an asymmetric half-bridge converter. Changing the operating mode according to the input voltage, the proposed converter can have low-voltage stresses on the switches and a low dc offset current of the transformer resulting in high efficiency for the entire input voltage range. In order to confirm the validity of the proposed converter, a 600 W prototype is implemented and experimented.

Published
2021
Full Text
View/download PDF

32. Highly improved passivation of c-Si surfaces using a gradient <italic>i</italic> a-Si:H layer.

Author

Lee, Soonil, Ahn, Jaehyun, Mathew, Leo, Rao, Rajesh, Zhang, Zhongjian, Kim, Jae Hyun, Banerjee, Sanjay K., and Yu, Edward T.

Subjects
*SOLAR cell efficiency, *INTEGRATED circuit passivation, *HETEROJUNCTIONS, *ELECTRIC properties of silicon, *INTERFACIAL bonding
Abstract

Surface passivation using intrinsic a-Si:H (i a-Si:H) films plays a key role in high efficiency c-Si heterojunction solar cells. In this study, we demonstrate improved passivation quality using i a-Si:H films with a gradient-layered structure consisting of interfacial, transition, and capping layers deposited on c-Si surfaces. The H2 dilution ratio (R) during deposition was optimized individually for the interfacial and capping layers, which were separated by a transition layer for which R changed gradually between its values for the interfacial and capping layers. This approach yielded a significant reduction in surface carrier recombination, resulting in improvement of the minority carrier lifetime from 1480 μs for mono-layered i a-Si:H passivation to 2550 μs for the gradient-layered passivation approach. [ABSTRACT FROM AUTHOR]

Published
2018
Full Text
View/download PDF

Catalog

Books, media, physical & digital resources
See catalog results