My research spans conformal inference, reliable machine learning, scientific AI, and interpretable modeling. For citation counts and the latest updates, visit Google Scholar or arXiv.
Accepted to NeurIPS 2026Conference paper
ORCA uses density-ranked generative samples and optimized coverage allocation to construct efficient prediction regions with finite-sample marginal coverage. Project page and interactive demo.
Cite the current arXiv version: Minxing Zheng and Shixiang Zhu. (2024). "Generative Conformal Prediction with Vectorized Non-Conformity Scores." arXiv:2410.13735.
Current arXiv preprint
JournalISPRS International Journal of Geo-Information, 2022
Interactive visualization and representation analysis for understanding glacier segmentation models, diagnosing data issues, and guiding model improvement.
Recommended citation: Minxing Zheng, Xinran Miao, and Kris Sankaran. (2022). "Interactive Visualization and Representation Analysis Applied to Glacier Segmentation." ISPRS International Journal of Geo-Information, 11(8), 415.
Download Paper
PreprintarXiv preprint
A prior-free conformal framework for constructing finite-sample valid uncertainty sets for latent distributional parameters using only observed context–response pairs and a specified forward model.
Recommended citation: Minxing Zheng, Wenbin Zhou, and Shixiang Zhu. (2026). "Beyond Predicting Responses: Conformal Inference for Latent Distributional Parameters." arXiv:2608.03607.
Download Paper
PreprintarXiv preprint
A physics-informed, test-oriented learning framework for detecting dynamical instability under distribution shift with controlled Type I error.
Recommended citation: Minxing Zheng, Zewei Deng, Liyan Xie, and Shixiang Zhu. (2026). "Learning to Test: Physics-Informed Representation for Dynamical Instability Detection." arXiv:2604.10967.
Download Paper
PreprintarXiv preprint
A defense against model extraction that preserves benign utility without relying on out-of-distribution query assumptions.
Recommended citation: Xueqi Cheng, Minxing Zheng, Shixiang Zhu, and Yushun Dong. (2025). "MISLEADER: Defending against Model Extraction with Ensembles of Distilled Models." arXiv:2506.02362.
Download Paper
PreprintarXiv preprint; revised 2026
A tuning-parameter-free, fully nonparametric framework for detecting and locating distributional changes in metric-space data using distance profiles.
Recommended citation: Paromita Dubey and Minxing Zheng. (2023; revised 2026). "Change Point Inference for Non-Euclidean Data Sequences using Distance Profiles." arXiv:2311.16025.
Download Paper