Hyeonsu Lee

Mr. Hyeonsu Lee

orcid

Research Focus & Metrics

Academic Career Length

7Years (2020-Present)

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

Representative Publications

View All Publications
5 September 2024 in AI
Verified
Journal Article
Open access
PDF file
HLHyeonsu Lee
Hyeonsu Lee
SHSeung-Hwan Hong
Seung-Hwan Hong
GKGwan-Heon Kim
Gwan-Heon Kim
HYHye Jin You
Hye Jin You
ELEun-Young Lee
Eun-Young Lee
JKJune Hyuk Kim
June Hyuk Kim
Technological advances in information-processing capacity have enabled integrated analyses (multi-omics) of different omics data types, improving target discovery and clinical diagnosis. This study proposes novel artificial intelligence (AI) learning strategies for incomplete datasets, common in omics research. The model comprises (1) a multi-omics generative model based on a variational auto-encoder that learns tumor genetic patterns based on different omics data types and (2) an expanded classification model that predicts cancer phenotypes. Padding was applied to replace missing data with virtual data. The embedding data generated by the model accurately classified cancer phenotypes, addressing the class imbalance issue (weighted F1 score: cancer type > 0.95, primary site > 0.92, sample type > 0.97). The classification performance was maintained in the absence of omics data, and the virtual data resembled actual omics data (cosine similarity mRNA gene expression > 0.96, mRNA isoform expression > 0.95, DNA methylation > 0.96). Meanwhile, in the presence of omics data, high-quality, non-existent omics data were generated (cosine similarity mRNA gene expression: 0.9702, mRNA isoform expression: 0.9546, DNA methylation: 0.9687). This model can effectively classify cancer phenotypes based on incomplete omics data with data sparsity robustness, generating omics data through deep learning and enabling precision medicine.

31 January 2021 in The Journal of Korean Institute of Information Technology
Verified
Journal Article
Closed access
HLHyeonsu Lee
Hyeonsu Lee
SHSeongeun Hong
Seongeun Hong
JBJunil Bang
Junil Bang
HKHwajong Kim
Hwajong Kim
고도로 발달된 개인 정보 식별 기술에 의해 개인에 대한 식별이 용이해지면서, 정보 사회에서 정보 주체의 권리를 보장할 수 있는 다양한 방안이 요구되고 있다. 연합학습은 이러한 요구에 의해 제안된 기계학습 방식으로 데이터를 비공개로 유지하면서 기계학습 알고리즘을 훈련하기 위한 특정 접근 방식이다. 본 논문에서는 개인정보보호 이슈에 민감한 의료 산업에 연합학습을 적용할 때 발생할 가능성이 있는 문제점을 파악하기 위해 망막 환자 데이터셋을 실제 의료기관이 데이터를 보유하고 있는 환경처럼 데이터 분포를 불균형하게 분할했다. 여기서 발생하는 클래스 불균형 문제에 다양한 학습 최적화 기법을 적용한 실험을 진행한 결과, 언더 샘플링 및 TopkAvg 기법을 적용한 실험에서 F1 score 0.96을 달성했으며, 학습 시간도 단축시켰다.

30 September 2020 in The Journal of Korean Institute of Information Technology
Verified
Journal Article
Closed access
HLHyeonsu Lee
Hyeonsu Lee
SHSeongeun Hong
Seongeun Hong
JBJunil Bang
Junil Bang
HKHwajong Kim
Hwajong Kim

14 November 2022 in JMIR Preprints
Verified
Preprint
Open access
HLHyeonsu Lee
Hyeonsu Lee
SHSeung Hwan Hong
Seung Hwan Hong
GKGwan Heon Kim
Gwan Heon Kim
JKJune Hyuk Kim
June Hyuk Kim
HYHye Jin You
Hye Jin You
JAJinwoo Ahn
Jinwoo Ahn
BACKGROUND Rapid technological advances have increased the diversity of omics data. In addition, technological advances and improvements in information-processing capacity have made integrated analyses (multi-omics) of different omics data types possible. Although multi-omics models are more advantageous for target discovery and clinical diagnosis than single-omics models, efficiency and performance are limited by data-handling steps, such as the need to change the model structure when a new omics data type is added and variation in available data among samples. OBJECTIVE This study proposes a novel artificial intelligence (AI) model and learning strategies for the use of incomplete datasets, which are common in omics research. The following goals were set: 1) to design a single AI model that can analyze both complete and incomplete data, 2) to design an AI model that can infer missing omics data by learning from partial omics data as input, and 3) to compare the proposed novel approach with previous methods that do not allow the use of incomplete data. METHODS The proposed model consists of two key components: (1) a multi-omics generative model based on the variational auto-encoder (VAE) that can learn genetic patterns in tumors based on different omics data types, and (2) an expanded classification model that can predict cancer phenotypes. In addition, padding was applied to replace missing omics data in each sample and generate virtual data. RESULTS The embedding data generated by the proposed model has three characteristics. First, its accuracy for classifying various cancer phenotypes (cancer type, sample type, and primary site) was high, addressing the class imbalance problem (cancer type weighted F1 score > 95%). Second, the virtual omics data generated by the model resembled the actual omics data (mean absolute error < 0.09). Third, the performance for classifying cancer phenotypes was higher for the model that could learn from incomplete data and complete data than for the model that could learn from complete data alone(primary site weighted F1 score: 0.0113 improvement). CONCLUSIONS The proposed novel model was capable of utilizing incomplete omics data. This showed good classification performance for cancer phenotypes and effective data construction in cases of missing omics data. Thus, overcoming data limitations, generating omics data through deep learning, and contributing to the realization of precision medicine. The model can be expanded to any omics data type for cancer, in addition to the data types evaluated in this study, suggesting that model performance can be further improved.