Zishi Zhang

Zishi Zhang

Peking University

Hi🙋! I am a PhD student at the Guanghua School of Management (GSM), Peking University. My research focuses on the interface between Operations Management (OM)/ Operations Research (OR), and AI.

📝 Publications

Selected Working Papers

  • Debiasing AI Simulations with Human Data. [SSRN] Jinhui Han, Ming Hu, Zishi Zhang (alphabetical order)

  • LLM-Inspired Pretrain-Then-Finetune for Small-Data, Large-Scale Optimization. [SSRN][arxiv] Zishi Zhang, Jinhui Han, Ming Hu, Yijie Peng

  • Nonparametric Bayesian Optimization for General Rewards. [arxiv] Zishi Zhang, Tao Ren, Yijie Peng

    • Under revision at Operations Research.
  • Optimal Low-Rank Stochastic Gradient Estimation for LLM Training. [arxiv] Zehao Li, Tao Ren, Zishi Zhang, Xi Chen, Yijie Peng
    • Revise and resubmit at Operations Research.
  • Sample-Efficient “Clustering and Conquer” Procedures for Parallel Large-Scale Ranking and Selection. [arxiv] Zishi Zhang, Yijie Peng
    • Under revision at Naval Research Logistics.

AI Conference Papers

* indicates equal contribution.

  • Half-order Fine-Tuning for Diffusion Model: A Recursive Likelihood Ratio Optimizer. ICLR 2026 Oral🔥🔥 Tao Ren, Zishi Zhang*, Zehao Li, Jingyang Jiang, Shentao Qin, Guanghao Li, Yan Li, Yi Zheng, Xinping Li, Min Zhan, Yijie Peng.
  • Exploring and Exploiting Model Uncertainty in Bayesian Optimization. NeurIPS 2025. Zishi Zhang, Tao Ren, Yijie Peng.
  • FLOPS: Forward Learning with OPtimal Sampling. ICLR 2025. Tao Ren, Zishi Zhang*, Jinyang Jiang, Guanghao Li, Zeliang Zhang, Mingqian Feng, Yijie Peng.
  • Dynamic Assortment Optimization in Live-Streaming Sales. WSC 2024. Zishi Zhang, Haidong Li, Ying Liu, Yijie Peng
  • RiskPO: Risk-based Policy Optimization with Verifiable Reward for LLM Post-Training. ICLR 2026. Tao Ren, Jinyang Jiang, Hui Yang, Wan Tian, Minhao Zou, Guanghao Li, Zishi Zhang, Qinghao Wang, Shentao Qin, Yanjun Zhao, Rui Tao, Hui Shao, Yijie Peng  

🚧 Work in progress

I am currently working on several projects at the intersection of generative AI, AI agents and decision-making.

  • Agentic AI for Operations Management and social science. Focusing on AI agents for OM applications, such as pricing and recommendation, as well as social science applications. My work has two main directions: (i) developing theoretical foundations to understand their limitations and identify potential failure modes; and (ii) designing optimization and debiasing methods to better align these agents with human preferences, values, and decision-making needs.

  • Large-Scale Stochastic Optimization for Foundation Models. Developing scalable, computationally efficient gradient-based and gradient-free optimization methods to train LLMs and multimodal generative models.

  • Generative models for decision-making. Developing tailored generative models for decision-making problems in operations management and finance, and establishing their theoretical foundations.

  • AI for Science. Applying AI methods to accelerate scientific discovery in chemistry and other scientific domains.

🏛️ Presentations

  • Phd Colloquium. Winter Simulation Conference 2023. San Antonio, Texas.
  • Large-scale simulation optimization. INFORMS Annual Meeting 2024. Seattle, Washington.
  • AI, Optimization, and Platform Decisions. POMS-HK 2026. Shen Zhen, China.
  • CSAMSE 2026. Shenyang, China.