Park, Si Hyung (Sihyung)
Ph.D. Student in Statistics
Department of Statistics
North Carolina State University
spark52 at ncsu dot edu
Brief Biography
I am Sihyung Park (pronounced “See-Young”), a first-year Ph.D. student in the Department of Statistics at North Carolina State University. Prior to joining the Ph.D. program, I received a B.S. in Life Sciences (thesis advisor: Prof. Sung Wook Chi) and a B.Econ. in Statistics (double major) from Korea University. After that, I received an M.S. in Statistics from Seoul National University (advisor: Prof. Byeong Uk Park).
I am interested in applying machine learning methods in a socially responsible way. Since most of the current issues result from the colossal, overparametrized structure of neural networks, my interest is in the compression of such models while retaining their performance. My current interests lie broadly in low-rank latent structure, tensor decomposition methods, dimension reduction, and nonparametric (shape-constrained) function estimation.
Education
- North Carolina State University (Sep. 2022 - Present)
- First-year Ph.D student in Statistics
- Seoul National University (Mar. 2020 - Feb. 2022)
- M.S. in Statistics
- Thesis: Large-factor compression of neural networks via tensor decomposition
- Korea University (Mar. 2014 - Feb. 2018)
- B.S. in Life Sciences
- B.Econ. in Statistics (double major)
- Thesis: Deconvolution of Ago CLIP-Seq resolves genome-wide miRNA target sites
Experiences
- Constructing intelligent water purification system (Sep. 2021 - Feb.2022)
- Comissioned by WIDEP and Korea Ministry of Environment
- Implemented a Bayesian contextual linear bandit with Thompson sampling
- Interval prediction for an undisclosed regression model (Apr. 2021 - Aug. 2021)
- Comissioned by Woori Bank
- Implemented a residual bootstrapping-based prediction interval
- Applied nonparametric isotonic concave regression to model uncertainty in time
- Estimation and forecasting of recall status of defective refrigerators (Mar. 2021 - Jan. 2022)
- Comissioned by Korean Agency for Technology and Standards
- Implemented an autoencoder-based anomaly detector
- Fitted piecewise logistic growth curve with press coverage frequency as a covariate
Publications
Peer-reviewed journal articles
- S. Park, S. H. Ahn, E. S. Cho, Y. K. Cho, E. S. Jang, S. W. Chi. (2018). “CLIPick: a sensitive peak caller for expression-based deconvolution of HITS-CLIP signals”. Nucleic acids research. 46(21), 11153-11168. (link)
Honors and Awards
- Korean Government Scholarship for Study Overseas - Korean Government (2022-2024)
- Academic Excellence Scholarship - Seoul National University (Fall 2020, Fall 2021)
- Summa Cum Laude - Korea University (Feb. 2018)
- Korea University Future Research Grant - Korea University (2016-2018)
- Academic Excellence Award - Korea University (2014-2015, 2017)
About the Blog
De Novo is a word frequently used in Biology, which means “of new”, or “from the beginning” in Latin. This blog is a journal written by a student who wants to make a living by studying and doing research. Contents are mostly about some interesting or important ideas I encountered while studying and doing research. This is my attempt to build up and organize new (and hopefully meaningful) knowledge.
Goni and Bori, my cockatiels. They love getting head scritches.
Resources