Wonhee Lee

Wonhee Lee

Auto-generated profile, based on publicly available information

Wonhee Lee

Department of Autonomous Manufacturing Research Center, Korea Electronics Technology Institute, Seongnam-si 13509, Republic of Korea

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 February 2025 in The International Journal of Advanced Manufacturing Technology
Verified
Journal Article
Open access
PDF file
YLYongho Lee
Yongho Lee
JSJunho Shin
Junho Shin
WLWonhee Lee
Wonhee Lee
This paper presents a data-driven approach to improving the productivity of manufacturing companies operating under Make-To-Order (MTO). In this study, a comprehensive analysis of processes, time, resources, and quality is performed using process mining techniques. This enables an understanding of the manufacturing process flow from a global perspective and addresses bottlenecks and workload issues from a local perspective in the manufacturing environment. This approach was implemented in a fully automated machining and logistics testbed developed by the Korea Electronics Technology Institute. Through a case study, the practical application and effectiveness of this approach are demonstrated, including specific improvement proposals. The validation of these proposals through simulations, focusing on key processes, resulted in significant productivity improvements. Ultimately, this study aims to build a more efficient and competitive manufacturing environment by showcasing the potential of process mining and various data visualization and analysis techniques. The results of this study demonstrate that adhering to the proposed framework enables continuous process optimization and improved operational performance is achievable in the manufacturing sector.

30 November 2025 in The Journal of Korean Institute of Information Technology
Verified
Journal Article
Closed access
SJseongsu jhang
seongsu jhang
DYDonghwi Yoo
Donghwi Yoo
JKJaeyoung Kwon
Jaeyoung Kwon
WLWonhee Lee
Wonhee Lee
생성형 인공지능의 발전과 제조 산업에서의 AI 도입 확산은 열악한 환경에서도 적용 가능한 고속 비지도 이상 탐지 시스템 수요를 높이고 있다. 특히, 비전 검사 AI 모델은 사람의 오류를 줄이고 속도를 높일 수 있으나, 실제 현장에서는 제품 수명 주기 단축과 다품종 소량 생산으로 인해 불량 데이터 확보와 라벨링에도 한계가 있다. 이에 따라, 정상 데이터만으로 학습 가능한 비지도 방식이 주목받고 있다. 본 연구에서는 실제 가전부품 제조 공정 현장에서 수집한 고해상도 이미지를 활용해, 대표적 비지도 모델인 f-AnoGAN에 구조적 유사도 지수(SSIM)를 손실함수로 적용하여 실험적으로 검증하였다. 그 결과, 모델 학습 시 SSIM 손실을 적절한 가중치로 Feature Matching 손실과 함께 적용할 경우 이상 탐지 성능이 유의미하게 향상됨을 확인했다.

30 April 2026 in The Journal of Korean Institute of Information Technology
Verified
Journal Article
Closed access
SJseongsu jhang
seongsu jhang
DYDonghwi Yoo
Donghwi Yoo
JKJaeyoung Kwon
Jaeyoung Kwon
WLWonhee Lee
Wonhee Lee

30 June 2025 in The Journal of Korean Institute of Information Technology
Verified
Journal Article
Closed access
SJseongsu jhang
seongsu jhang
DYDonghwi Yoo
Donghwi Yoo
JKJaeyoung Kwon
Jaeyoung Kwon
WLWonhee Lee
Wonhee Lee
대규모 언어모델의 고도화 및 확산에 따라 한국어가 적용된 언어모델 또한 다양한 분야에서 적용, 활용되고 있다. 이는 다양한 서비스 확장과 산업 응용 가능성을 증대시키고 있으며, 특히 전문 지식이 요구되는 분야에서 정보 제공 및 지원 도구로서의 역할이 점차 중요해지고 있다. 하지만 숙련요소의 데이터화는 전문가의 개입이 필요하며 많은 작업 공수가 필요하다. 본 연구에서는 HITL(Human-in-the-Loop) 접근 방식을 활용하여 정밀 금속가공 업종에서의 전문 지식 및 숙련공의 암묵지와 노하우를 효과적으로 모델에 반영하는 방법을 제안한다. 언어모델을 파인튜닝하고, 전문가의 반복적 검증과 피드백을 통해 모델 성능을 최적화하여 가공 공정 관련 가이드를 위한 파인튜닝 기반 챗봇용 언어모델 학습을 연구하였다. 본 연구를 통하여 언어모델 파인튜닝에서의 HITL 적용 시 모델의 특성에 따른 비언어적 특성을 포함한 데이터 학습 성능을 검증하였다.

14 April 2026 in Materials
Verified
Journal Article
Open access
PDF file
SJseongsu jhang
seongsu jhang
JKJae-Young Kwon
Jae-Young Kwon
WLWon-Hee Lee
Won-Hee Lee
Hong-Gyu ParkHong-Gyu Park
Hong-Gyu Park
This study addresses energy-response prediction in sheet metal forming under small-data conditions, where conventional simulation-based approaches are computationally expensive and data acquisition is limited. We propose an Energy-Informed Neural Network (EINN) framework that integrates energy consistency constraints and a Fenchel–Young duality-based loss to enforce physically consistent learning without relying on explicit governing equations. Using a dataset generated from 54 finite element simulations across 18 materials and three friction conditions, the proposed model demonstrates significant performance improvements. Specifically, EINN achieves an RMSE of 0.0096, MAE of 0.0065, and R2 of 0.9778, corresponding to approximately a 48% reduction in RMSE compared to the best baseline model. Compared to an energy-constrained neural network without the Fenchel–Young term, prediction error is reduced by approximately 50% with substantially improved stability. These results indicate that embedding energy-consistent dual structures enhances both prediction accuracy and robustness, providing a practical surrogate modeling approach for process optimization in sheet metal forming under limited data availability.

28 August 2024 in Research Square
Verified
Preprint
Open access
PDF file
YLYongho Lee
Yongho Lee
JSJunho Shin
Junho Shin
WLWonhee Lee
Wonhee Lee
This paper presents a data-driven approach to improving the productivity of manufacturing companies operating under Make-To-Order (MTO). In this study, a comprehensive analysis of processes, time, resources, and quality is performed using process mining techniques. This enables an understanding of the manufacturing process flow from a global perspective and addresses bottlenecks and workload issues from a local perspective in the manufacturing environment. This approach was implemented in a fully automated machining and logistics testbed developed by the Korea Electronics Technology Institute. Through a case study, the practical application and effectiveness of this approach are demonstrated, including specific improvement proposals. The validation of these proposals through simulations, focusing on key processes, resulted in significant productivity improvements. Ultimately, this study aims to build a more efficient and competitive manufacturing environment by showcasing the potential of process mining and various data visualization and analysis techniques. The results of this study demonstrate that adhering to the proposed framework enables continuous process optimization and improved operational performance are achievable in the manufacturing sector.