Jian Xu

Jian Xu

he

About

I am Jian Xu, a postdoctoral researcher at RIKEN iTHEMS & AIP (Tokyo, Japan), hosted by Prof. Qibin Zhao. I received my Ph.D. in Computational Mathematics and Electronic Information from South China University of Technology in December 2025, under the supervision of Prof. Delu Zeng.

My research lies at the intersection of probabilistic machine learning, stochastic processes, generative modeling, differential-equation methods, and quantum machine learning. I focus on Bayesian generative models and principled inference, with particular interest in Gaussian / Student-t processes, variational inference, and diffusion models.

During my postdoc I work on quantum machine learning — a joint RIKEN iTHEMS / AIP effort to bring quantum methods into modern ML.

My first-author work has appeared in ICML (Oral, Top 1%), ICLR, AAAI, AISTATS, IEEE TNNLS, and KBS. I collaborate closely with Prof. John Paisley (Columbia University) on Bayesian machine learning and variational inference. Several manuscripts are currently under review.

🎓 Education

  • Ph.D. in Computational Mathematics & Electronic Information, South China University of Technology, 2020 – 2025
    • First-author work at top venues incl. ICML 2024 Oral (Top 1%) and ICLR 2026
    • Advisor: Prof. Delu Zeng · Hosts: Prof. Qibin Zhao & Prof. John Paisley
  • M.Sc. in Basic Mathematics, South China University of Technology, 2018 – 2020
    • First-class academic scholarship · Advisor: Assoc. Prof. Bingsheng Lin
  • B.A. in Business Administration, Communication University of China, 2013 – 2017
    • Second Prize, National College Student Mathematics Competition

🔬 Research Topics

  • Probabilistic machine learning
  • Deep Gaussian processes
  • Student-t processes & heavy-tailed modeling
  • Stochastic differential equations
  • Generative models & diffusion-based inference
  • Variational inference at scale
  • Quantum machine learning

📝 Selected Publications

A full list with paper PDFs and code is on the Publications page.

  • Diffusion Bridge Variational Inference for Deep Gaussian Processes — ICLR 2026 · paper · code
  • Sparse Variational Student-t Processes for Heavy-Tailed Modeling — IEEE TNNLS 2026 · paper · code
  • Bayesian Gaussian Process ODEs via Double Normalizing Flows — AISTATS 2025 · paper · code
  • Fully Bayesian Differential Gaussian Processes through SDEs — KBS 2025 · paper · code
  • Neural Operator Variational Inference for Deep GPs — IEEE TNNLS 2025 · paper · code
  • Sparse Inducing Points in Deep Gaussian Processes: DDVI — ICML 2024 (Oral, Top 1%) · paper · code
  • Sparse Variational Student-t Processes — AAAI 2024 · paper · code

📬 Contact