I am an assistant professor at the City University of Hong Kong. I got my Ph.D. from the Department of Statistics, Oxford, supervised by Prof. Tom Rainforth and Prof. Yee Whye Teh. Before that, I got my BS and MS degrees from Peking University and worked as a researcher at Bytedance AI lab. I lead the Miaow Lab. Our research focuses on machine reasoning (LLM reasoning, AI4Math) and generative models. For visit our group homepage for more information about our research projects and other group members.

---

News

* June 2026: We are awarded the ECS grant from Hong Kong Research Grants Council!
* May 2026: SSAE is accepted by ICML26. Congrats to Xuan, Jiayu, and Yuhang!

---

Publications

• TD-Grokking: Learning from Zero-Reward Problems by Training-Time Decomposition
  Ningyuan Xi, Hao Xu, Hongsheng Xin, Ning Miao
  [pdf] [code]
• Beyond Ideal Instruction: A Comprehensive Framework for Evaluating LLMs in Realistic Interactions
  Xuan Yang, Hao Xu, Tingfeng Hui, Hongsheng Xin, Kaike Zhang, Chunxiao Liu, Ning Miao
  [pdf] [code]
• STT-Arena: A More Realistic Environment for Tool-Using with Spatio-Temporal Dynamics
  Tingfeng Hui, Hao Xu, Pengyu Zhu, Hongsheng Xin, Kun Zhan, Sen Su, Chunxiao Liu, Ning Miao
  [pdf] [code] [dataset and model]
• Verifier-Backed Hard Problem Generation for Mathematical Reasoning
  Yuhang Lai, Jiazhan Feng, Yee Whye Teh, Ning Miao
  [pdf]
• Step-Level Sparse Autoencoder for Reasoning Process Interpretation
  Xuan Yang, Jiayu Liu, Yuhang Lai, Hao Xu, Zhenya Huang, Ning Miao
  In ICML, 2026.
  [pdf] [code]
• Not All Steps are Informative: On the Linearity of LLMs' RLVR Training
  Tianle Wang, Zhongyuan Wu, Shenghao Jin, Hao Xu, Wei Chen, Ning Miao
  [pdf] [code]
• Enhancing Large Language Model Reasoning with Reward Models: An Analytical Survey
  Qiyuan Liu, Hao Xu, Xuhong Chen, Wei Chen, Yee Whye Teh, Ning Miao
  [pdf]
• MARCOS: Deep Thinking by Markov Chain of Continuous Thoughts
  Jiayu Liu, Zhenya Huang, Anya Sims, Enhong Chen, Yee Whye Teh, Ning Miao
  [pdf]
• BED-LLM: Intelligent Information Gathering with LLMs and Bayesian Experimental Design
  Deepro Choudhury, Sinead Williamson, Adam Goliński, Ning Miao, Freddie Bickford Smith, Michael Kirchhof, Yizhe Zhang, Tom Rainforth
  In ICLR, 2026.
  [pdf]
• SelfCheck: Using LLMs to Zero-Shot Check Their Own Step-by-Step Reasoning
  Ning Miao, Yee Whye Teh, Tom Rainforth
  In ICLR, 2024.
  [bib] [pdf] [code]
• Learning Instance-Specific Augmentations by Capturing Local Invariances
  Ning Miao, Tom Rainforth, Emile Mathieu, Yann Dubois, Yee Whye Teh, Adam Foster, Hyunjik Kim
  In ICML, 2023.
  [bib] [pdf] [code]
• On Incorporating Inductive Biases into VAEs
  Ning Miao, Emile Mathieu, Siddharth N, Yee Whye Teh, Tom Rainforth
  In ICLR, 2022.
  [bib] [pdf] [code]
• Do You Have the Right Scissors? Tailoring Pre-trained Language Models via Monte-Carlo Methods
  Ning Miao, YuXuan Song, Hao Zhou, Lei Li
  In ACL, 2020.
  [bib] [pdf] [code]
• Dispersed Exponential Family Mixture VAEs for Interpretable Text Generation
  Wenxian Shi, Hao Zhou, Ning Miao, Shenjian Zhao, Lei Li
  In ICML, 2020.
  [bib] [pdf] [code]
• Improving Maximum Likelihood Training for Text Generation with Density Ratio Estimation
  YuXuan Song, Ning Miao, Hao Zhou, Lei Li
  In AISTATS, 2020.
  [bib] [pdf]
• Kernelized Bayesian Softmax for Text Generation
  Ning Miao, Hao Zhou, Chengqi Zhao, Wenxian Shi, Lei Li
  In NeurIPS, 2019.
  [bib] [pdf] [code]
• Generating Fluent Adversarial Examples for Natural Languages
  Huangzhao Zhang, Hao Zhou, Ning Miao, Lei Li
  In ACL, 2019.
  [bib] [pdf] [code]
• Constrained Sentence Generation via Metropolis-Hastings Sampling
  Ning Miao, Hao Zhou, Lili Mou, Rui Yan, Lei Li
  In AAAI, 2019.
  [bib] [pdf] [code]

---

Technical Blogs

• [23/June/2025] Entropy Maximization Alone Can Improve LLM Reasoning Performance? [Notion]

---

This webpage was built with Bootstrap and Jekyll. You can find the source code here. Last updated: Jun 29, 2026