GeezerVison

I am creating continually self-improving AI systems. Currently, I am a Member of Technical Staff at OpenAI. I obtained my Ph.D. from Stanford University advised by Emmanuel Candès and Tatsunori Hashimoto, and my B.A. from Berkeley with Yi Ma and Jacob Steinhardt.

---

Talks

• PhD thesis defense [slides]
• Stanford CS525 guest lecture on synthetic data [slides]
• Stanford statistics seminar on synthetic continued pretraining [slides]

---

Education

• Ph.D. Statistics, Stanford University (2022-2025)
• M.S. Electrical Engineering, University of California, Berkeley (2020-2021)
• B.A. Mathematics, Highest Honors, University of California, Berkeley (2016-2020)
• B.A. Computer Science, University of California, Berkeley (2016-2020)
• B.A. Physics, University of California, Berkeley (2016-2020)

---

Research

• Towards execution-grounded automated AI research
  Chenglei Si*, Zitong Yang*, Yejin Choi, Emmanuel Candès, Diyi Yang, Tatsunori Hashimoto
  International Conference on Machine Learning (ICML), 2026.
  [arXiv] [code]


• Synthetic bootstrapped pretraining
  Zitong Yang*, Aonan Zhang*, Hong Liu, Tatsunori Hashimoto, Emmanuel Candès, Chong Wang, Ruoming Pang
  International Conference on Learning Representations (ICLR), 2026.
  [arXiv] [slides]


• s1: simple test-time scaling
  Niklas Muennighoff*, Zitong Yang*, Weijia Shi*, Xiang Lisa Li*, Li Fei-Fei, Hannaneh Hajishirzi, Luke Zettlemoyer, Percy Liang, Emmanuel Candès, Tatsunori Hashimoto
  Empirical Methods in Natural Language Processing (EMNLP, Oral), 2025.
  [arXiv] [code] [slides] [slides(Chinese)]


• Synthetic continued pretraining
  Zitong Yang*, Neil Band*, Shuangping Li, Emmanuel Candès, Tatsunori Hashimoto
  International Conference on Learning Representations (ICLR, Oral), 2025.
  [arXiv] [code] [slides] [slides(Chinese)]


• Bellman conformal inference: calibrating prediction intervals for time series
  Zitong Yang, Emmanuel Candès, Lihua Lei
  arXiv preprint, 2024.
  [arXiv] [code]


• ResMem: Learn what you can and memorize the rest
  Zitong Yang, Michal Lukasik, Vaishnavh Nagarajan, Zonglin Li, Ankit Singh Rawat, Manzil Zaheer, Aditya Krishna Menon, Sanjiv Kumar
  Neural Information Processing Systems (NeurIPS), 2023.
  [arXiv] [slides]


• Predicting out-of-distribution error with the projection norm
  Yaodong Yu*, Zitong Yang*, Alexander Wei, Yi Ma, Jacob Steinhardt.
  International Conference on Machine Learning (ICML), 2022.
  [arXiv] [slides]


• Exact gap between generalization error and uniform convergence in random feature models
  Zitong Yang, Yu Bai, Song Mei.
  International Conference on Machine Learning (ICML), 2021.
  [arXiv] [slides]


• Understanding generalization in adversarial training via the bias-variance decomposition
  Yaodong Yu*, Zitong Yang*, Edgar Dobriban, Jacob Steinhardt, Yi Ma.
  arXiv preprint, 2021.
  [arXiv]


• Rethinking the Bias-variance dilemma for generalization of neural networks
  Zitong Yang*, Yaodong Yu*, Chong You, Jacob Steinhardt, Yi Ma.
  International Conference on Machine Learning (ICML), 2020.
  [arXiv] [slides]


• Complete dictionary learning via l4-norm maximization over the orthogonal group.
  Yuexiang Zhai, Zitong Yang, Zhenyu Liao, John Wright, Yi Ma
  Journal of Machine Learning Research (JMLR), 2020.
  [arXiv]