Jongmin Mun

Email | Google Scholar
Twitter | LinkedIn

I am a third-year PhD student in the Data Sciences and Operations Department at the University of Southern California. I am advised by Prof. Yingying Fan and Prof. Paromita Dubey. My current research focuses on using bandit algorithms as an optimization tool to address problems in statistics and operations research, including high-dimensional clustering and dynamic pricing.

Before starting my PhD, I studied:

Privacy–utility trade-off in private federated anayltics (A/B testing) using minimax statistical theory with Prof. Ilmun Kim,
High-dimensional neural signal analysis for the development of liquid-metal neural probe, with Dr. Young-Geun Park and Prof. Jang-Ung Park,
Regression for mixed functional-Euclidean data for medical image analysis, with Prof. Jeong Hoon Jang,
The use of generative models to address class imbalance in machine learning, with Prof. Jaeoh Kim, at the Center for Army Analysis and Simulations (CAAS), a research department of the Republic of Korea Army.

News

Jan 25, 2026	A new paper for simultaneous supervised dimension reduction and regression in multimodal data is now on arxiv. Motivated by renal imaging studies (combining renogram curves with covariates), the approach is designed to effectively handle intermodal correlations.
Dec 6, 2025	I presented my dynamic pricing paper at NeurIPS MLxOR workshop. Under the covariate shift and local differential privacy, we use pessimistic value-based off-policy learning algorithm for continuous treatment (pricing) assignment policy.
Nov 10, 2025	My paper has been published in the Journal of Machine Learning Research! 🎉 We propose private permutation-based two-sample tests incorporating U-statistic and Google’s RAPPOR mechanism, and we derive the minimax rates that characterize the privacy–utility trade-off.
May 26, 2025	My first paper as a PhD student is now on arXiv! We propose an efficient iterative high-dimensional clustering algorithm that bypasses model parameter estimation.
Nov 24, 2024	My collaboration on detecting sex differences in autism using brain connectome data is accepted by NeuroImage! I contributed by using a generative model to enhance testing power.

Selected Publications

Privacy

Minimax optimal two-sample testing under local differential privacy

Jongmin Mun, Seungwoo Kwak, and Ilmun Kim

Journal of Machine Learning Research, 2025

Summary Published Version Slides Python Package

Federated learning, where user data remains on local devices and only weight updates are transmitted, has become the industry standard for training neural networks on sensitive data. Federated analytics builds on this infrastructure, enabling data science insights without transmitting raw data. However, recent studies show that hijacking weight updates or data summaries can allow recovery of the original raw data. To address this privacy risk, we propose a private A/B testing method that transmits noisy data summaries within the federated learning framework. We begin with multinomial data, introducing private permutation tests using privacy mechanisms like the Laplace mechanism, discrete Laplace mechanism, and Google’s RAPPOR mechanism. We extend our approach to continuous data using binning and analyze uniform separation rates under local differential privacy (LDP). Our tests rigorously control Type I error, satisfy LDP constraints, and achieve minimax separation rates, highlighting the inherent privacy-utility trade-offs in private testing.

Imbalance

Weighted support vector machine for extremely imbalanced data

Jongmin Mun, Sungwan Bang, and Jaeoh Kim

Computational Statistics & Data Analysis, Mar 2025

Summary Published Version

Data augmentation and loss function adjustments are two common techniques for imbalanced classification. In cases of extreme class imbalance, it is often standard practice to combine these two approaches. However, determining the optimal oversampling ratio and the degree of asymmetry in the loss function typically relies on heuristics. We further consider the scenario where the minority class consists of subgroups and propose a straightforward method for combining data augmentation and loss function adjustments. This method serves as a sample-based approximation of the population-level asymptotically Bayes optimal oracle procedure.

Hi-dim

Hybrid Partial Least Squares Regression with Multiple Functional and Scalar Predictors

Jongmin Mun, and Jeong Hoon Jang

arxiv preprint, Jan 2026

Summary Preprint

Motivated by renal imaging studies that combine renogram curves with pharmacokinetic and demographic covariates, we propose Hybrid partial least squares (Hybrid PLS) for simultaneous supervised dimension reduction and regression in the presence of cross-modality correlations. The proposed approach embeds multiple functional and scalar predictors into a unified hybrid Hilbert space and rigorously extends the nonlinear iterative PLS (NIPALS) algorithm. This theoretical development is complemented by a sample-level algorithm that incorporates roughness penalties to control smoothness. By exploiting the rank-one structure of the resulting optimization problem, the algorithm admits a computationally efficient closed-form solution that requires solving only linear systems at each iteration. We establish fundamental geometric properties of the proposed framework, including orthogonality of the latent scores and PLS directions. Extensive numerical studies on synthetic data, together with an application to a renal imaging study, validate these theoretical results and demonstrate the method’s ability to recover predictive structure under intermodal multicollinearity, yielding parsimonious low-dimensional representations.
High-Dimensional Sparse Clustering via Iterative Semidefinite Programming Relaxed K-Means

Jongmin Mun, Paromita Dubey, and Yingying Fan

arxiv preprint, May 2025

Summary Slides Preprint

We propose an iterative algorithm for clustering high-dimensional data, where the true signal lies in a much lower-dimensional space. Our method alternates between feature selection and clustering, without requiring precise estimation of sparse model parameters. Feature selection is performed by thresholding a rough estimate of the discriminative direction, while clustering is carried out via a semidefinite programming (SDP) relaxation of K-means. In the isotropic case, the algorithm is motivated by the minimax separation bound for exact recovery of cluster labels using varying sparse subsets of features. This bound highlights the critical role of variable selection in achieving exact recovery. We further extend the algorithm to settings with unknown sparse precision matrices, avoiding full model parameter estimation by computing only the minimally required quantities. Across a range of simulation settings, we find that the proposed iterative approach outperforms several state-of-the-art methods, especially in higher dimensions.

Brain

In-vivo integration of soft neural probes through high-resolution printing of liquid electronics on the cranium

Young-Geun Park, Yong Won Kwon, Chin Su Koh, Enji Kim, Dong Ha Lee, Sumin Kim, Jongmin Mun, Yeon-Mi Hong, Sanghoon Lee, Ju-Young Kim, Jae-Hyun Lee, Hyun Ho Jung, Jinwoo Cheon, Jin Woo Chang, and Jang-Ung Park

Nature Communications, Feb 2024

Summary Published Version

To enable high-quality neural recordings in freely moving animals, our materials science team developed a liquid metal–based neural probe that can be directly printed onto the brain, along with elastic electronic circuits. To validate the fidelity of the signals recorded by this neural probe, we statistically assessed the long-term stability of signal clustering and examined the correlations between micro-scale neuronal activity and macro-level brain dynamics during mouse behavioral experiments involving visual and motor stimuli.