Projects | Jongmin Mun

Validating the signal fidelity of novel soft neural probes

Nature Communications

Filtering, feature extraction, clustering, regression and statistical testing of neural data.

Causal inference of sex differences in autism spectrum disorder

NeuroImage

Stabilizing heterogeneous treatment effect estimation for high-dimensional outcomes in high-imbalance regimes

fMRI Site Harmonization Using Causal Inference

working paper

Site Harmonization Using Average Treatment Effect for the Overlap Population

Federated A/B Testing under User-level Privacy Protection

Journal of Machine Learning Research

Nonparametric two-sample testing under local differential privacy.

Offline Dynamic Pricing under Covariate Shift and Local Differential Privacy via Twofold Pessimism

NeurIPS 2025 MLxOR Workshop

We propose a DM (direct method)-type transfer learning algorithm for learning continuous treatment assignment policy from offline data under local differential privacy.

Predicting Kidney Obstruction from Renal Photon Count Curve and Auxiliary Variables

Under revision

We propose a computationally efficient NIPALS-type hybrid partial least squares regression algorithm that unifies functional and scalar predictors in a Hilbert space.

Off-Policy Learning in Partially Observed Markov Decision Processes under Sequential Ignorability

Working Paper

We stabilize off-policy learning from partially observed health data using a geometry-aware gradient method, improving convergence and policy performance in simulations.

Wildfire Prediction from Imbalanced Data

Computational Statistics & Data Analysis Journal of Classification

Predicting forest fire from artillery training data and weather data

Private and Scalable Hash Kernel Bit-flip Two-Sample Testing

Working Paper

We propose using Hash Kernels to compress high-dimensional data before bit-flip privatization, making locally private A/B testing viable for massive, unknown alphabets.

Bayesian Spatial Modeling of Cancer Incidence (BYM2 Model)

Based on ch 15, 16 of BDA3, we run Bayesian GLM with spatial random effect on Manhattan breast cancer incidence.

Operations Management

Coursework and projects related to Operations Management and Optimization.