About me

Hi, I am Jinwon Sohn from South Korea and currently I am a Principal Researcher at the University of Chicago Booth under the supervision of Prof. Veronika Rockova. I earned my Ph.D. in Statistics at Purdue University under the supervision of Prof. Qifan Song and master degree at Yonsei University under Prof. Taeyoung Park. Before entering the Ph.D. program, I worked at Datarize as a data scientist, contributing to evaluating recommendation models.

My research asks how statistical guarantees – valid inference, training stability, fairness, and privacy – can be established when data are generated, transformed, or constrained by algorithms such as generative AI. My doctoral work built the generation-side foundation for this question through fairness-aware learning, stable generative modeling, and synthetic-data release, with work appearing in the Journal of the American Statistical Association (JASA), IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), and the International Conference on Artificial Intelligence and Statistics (AISTATS). My current and future work moves to the inferential side, asking, once data are synthetic, privacy-protected, or AI-generated, what conclusions can a statistician trust?

Research Interests 💡

Statistical foundations for synthetic and AI-generated data; Trustworthy statistical machine learning; modern Bayesian inference

News ⭐

[June 2026] Sohn, J. and Song, Q. Parallelly Tempered Generative Adversarial Nets: Toward Stabilized Gradients. (2026+). Accepted in Journal of American Statistical Association (in-press).
[May 2026] Sohn, J., Song, Q., & Lin, G. Task-tailored Pre-processing: Fair Downstream Supervised Learning. (2026+). Accepted in Transactions on Pattern Analysis and Machine Intelligence (in-press).
[May 2025] Lim, T., Nam, K., and Sohn, J. Monotone Curve Estimation with Convex Duality. Under Major Revision in Journal of American Statistical Association.
[Oct 2025] My daughter Yena Sohn was born! 참 고생 많았다 Sangmi!
[July 2025] Jinwon Sohn joined Chicago Booth as a postdoc under the supervision of Prof. Veronika Rockova.

Recent Publications 📚

Sohn, J., & Song, Q. (2026+). Parallelly Tempered Generative Adversarial Nets: Toward Stabilized Gradients. Journal of the American Statistical Association (in-press).
Sohn, J., Song, Q., & Lin, G. (2026+). Task-tailored Pre-processing: Fair Downstream Supervised Learning. Transactions on Pattern Analysis and Machine Intelligence (in-press).
Sohn, J., Song, Q., & Lin, G. (2024). Fair Supervised Learning with A Simple Random Sampler of Sensitive Attributes. AISTAT (pp. 1594-1602). PMLR.
Kang, T., Kim, S., Sohn, J., & Awan, J. (2024). Differentially Private Topological Data Analysis. Journal of Machine Learning Research.
Sohn, J., Jeong, S., Cho, Y. M., & Park, T. (2023). Functional Clustering Methods for Binary Longitudinal Data with Temporal Heterogeneity. Computational Statistics & Data Analysis, 185, 107766.