About

Hi! I am Soheun Yi, a second-year PhD student at Department of Statistics & Data Science, Carnegie Mellon University.

My research interest spans theoretical aspects of machine learning, somewhere between optimization theory and statistics. I am particularly fascinated by those topics that are inspired by real-world applications.

I am working on applications of statistics and machine learning to physical sciences. Before starting my PhD, I was a research scientist at Krafton AI. During my undergraduate, I was fortunate to be advised by Professor Ernest K. Ryu and Sanghack Lee. I started my research journey on theoretical aspects of machine learning, mainly on optimization theory and graphical models.

I was a quantitative analyst in Hyperithm, where I built a solid background in data analysis, especially in processing and analyzing large datasets.

You can find my CV here.

Contact

Email: soheuny [at] andrew [dot] cmu [dot] edu

News

• 2024.12. I will be presenting our work entitled “Toward Model-Agnostic Detection of New Physics Using Data-Driven Signal Regions” at ML4PS Workshop at NeurIPS 2024.
• 2024.06. A paper entitled “Convergence Analyses of Davis-Yin Splitting via Scaled Relative Graphs” has been accepted to SIAM Journal on Optimization (SIOPT).
• 2024.01. A paper entitled “Filter, Rank, and Prune: Learning Linear Cyclic Gaussian Graphical Models” has been accepted in AISTATS 2024.