Xiaojing Sun

About

I am a Ph.D. candidate in Statistics at Purdue University, advised by Prof. Fei Xue and Prof. Jianxi Su. I also maintain a long-term research collaboration with Prof. Xiao Wang. Prior to Purdue, I received my bachelor's degree from the School of Economics at Xiamen University in 2022, where I was advised by Prof. Jingyuan Liu.

I work on uncertainty quantification, finite-sample guarantees, risk control, and interpretability across complex data settings (e.g., wearable data, tail-risk modeling), and have recently been bringing these ideas to frontier AI research to make AI systems reliable and safe.

News

  • 2026/09

    New preprint: Which Self-Improvements Should We Trust? arXiv

  • 2026/06

    Started an internship at a private fund startup in Shanghai as a Quant Researcher, working with Kun Lu.

  • 2026/06

    New preprint: Statistical Proof as a Window into Human-AI Collaboration. arXiv

  • 2025/08

    Successfully completed my internship at Novartis (AMDS).

  • 2025/06

    ICSA Jiann-Ping Hsu Pharmaceutical and Regulatory Sciences Student Paper Award.

  • 2025/05

    Passed the preliminary exam and became an official Ph.D. candidate.

  • 2025/03

    George Casella Award (Purdue University).

  • 2024/08

    Honorable Mention in JSM Student Paper Award Competition (ASA).

Research and Projects

Research Paper

  1. Which Self-Improvements Should We Trust? Reliable Self-Improvement When Agents Reuse Their Benchmarks

    Xiaojing Sun, Yuhan Zeng, Zihua She, Xiao Wang.

    Risk-controlled evaluation of self-improving agents under adaptive benchmark reuse.

    2026Under reviewarXiv

  2. Statistical Proof as a Window into Human-AI Collaboration: Practical Insights and a Community Agenda

    Xiaojing Sun, Huayu Tang, Buxin Su, Mateo Matijasevick, Chong Wu, Fei Xue, Bingxin Zhao.

    Research-level statistical proofs as a lens on effective human–AI collaboration.

    2026arXiv

  3. Generalized Heterogeneous Functional Model with Applications to Large-scale Mobile Health Data

    Xiaojing Sun, Bingxin Zhao, Fei Xue.

    Personalized functional effects and subgroup discovery for wearable health data.

    Under revisionarXiv

  4. Flexible Non-exchangeable Bivariate Copulas for Mono-vine Structures: Construction, Angular Tail Dependence, and Applications

    Xiaojing Sun, Lei Hua, Jianxi Su, Zhongyi Yuan.

    Flexible modeling of tail dependence, asymmetry, and non-exchangeability with Pareto-mixture copulas.

    2025Under review

  5. TabCAM: Online Calibrated Regularization for Deep Learning

    Xiaojing Sun, Jianxi Su, Xiao Wang.

    Online calibration for adaptive regularization and built-in conditional uncertainty quantification.

    2026Under review

Selected Projects

Ongoing

An AI-Empowered Unified Framework for Leveraging Wearable Data in Accelerated Underwriting

Xiaojing Sun, Angela Gong, Jianxi Su, Xiao Wang, Fei Xue.

Integrating wearable signals into mortality models to quantify mortality slippage and inform accelerated life insurance underwriting.

In collaboration with Munich Re.

Ongoing

SkillForge-BE: An Agentic Harness for Adaptive Bioequivalence

Xiaojing Sun, Yuhan Zeng, Xiao Wang.

An agentic framework for adaptive bioequivalence analysis with a self-expanding skill library and anytime-valid inference constraints.

In collaboration with Eli Lilly.

Ongoing

Chat with GeoGebra

Xiaojing Sun, Ivory Zhang.

An AI agent that turns natural language into interactive GeoGebra visualizations; piloting in high schools in China.

Live demo

Professional Experience

  • Biostatistician Intern, Novartis (East Hanover, NJ) — Jun–Aug 2025
    Advanced Methodology & Data Science (AMDS). Mentors: Dr. Yao Chen (SM, AMDS) and Dr. Sophie Sun (AEA, AMDS). Manager: Dr. David Ohlssen (AEA, AMDS).
    Extended conditional inference trees to mixed-type multiple endpoints for prognostic and predictive variable selection, with family-wise error control; applied the framework to a psoriasis clinical trial.
  • Teaching & Research Assistant, Purdue Statistics — Aug 2022–Present

Awards

  • ICSA Jiann-Ping Hsu Pharmaceutical and Regulatory Sciences Student Paper Award (06/2025)
  • George Casella Award, Purdue University (03/2025)
  • Honorable Mention, JSM Student Paper Award Competition (08/2024)
  • National Scholarship, Ministry of Education of China (09/2020)
  • Finalist Prize, MCM/ICM (03/2020)

Talks / Presentations

  • Science VC of Novartis (East Hanover, NJ) — Internal Seminar (08/2025)
  • Actuarial Research Conference (Toronto, Canada) — Invited Talk (08/2025)
  • ICSA 2025 (Storrs, CT) — Invited Talk (06/2025)
  • JSM 2024 (Portland, OR) — Contributed Talk (08/2024)

Teaching

  • STAT 301 — Elementary Statistical Methods
  • STAT 475 — Intermediate Statistical Methodology
  • STAT 525 — Advanced Statistical Methods
  • STAT 526 — Advanced Statistical Methods

Contact

Email: sun1131@purdue.edu

Office: MATH 519
150 N. University St
West Lafayette, IN 47907

Last updated: