Han Cui

崔涵

I am a PhD in the WestlakeNLP Group at Westlake University and Zhejiang University, advised by Yue Zhang.

Previously, I earned both my bachelor's and master's degrees in Computer Science from the Harbin Institute of Technology.

My research focuses on Natural Language Processing and language agent learning. Previously, I studied how to improve model inference through control over intermediate reasoning processes, including explanation graph generation (EG3P), execution-based reasoning over heterogeneous data (HProPro), and value-guided policy optimization from a Q-learning perspective (DVO). My current goal is to develop training objectives that leverage process signals, such as partial solutions, tool-use traces, verifier feedback, and successful or failed action sequences, for more stable and sample-efficient agent learning in realistic digital environments.

From June to Sep 2023, I interned at Tencent PCG as an algorithm engineer. You can find my latest CV here.

News

Selected Publications

Deep Literature Survey Automation with an Iterative Workflow

Hongbo Zhang*, Han Cui*, Yidong Wang, Yijian Tian, Qi Guo, Cunxiang Wang, Jian Wu, Chiyu Song, Yue Zhang

arXiv preprint · Paper · Code

Direct Value Optimization: Improving Chain-of-Thought Reasoning in LLMs with Refined Values

Hongbo Zhang*, Han Cui*, Guangsheng Bao*, Linyi Yang, Jun Wang, Yue Zhang

EMNLP 2025 · Paper · Code

HProPro: Exploring Hybrid Question Answering via Program-based Prompting

Qi Shi*, Han Cui*, Haofeng Wang, Qingfu Zhu, Wanxiang Che, Ting Liu

ACL 2024 · Paper · Code

Explanation Graph Generation via Generative Pre-training over Synthetic Graphs

Han Cui, Shangzhan Li, Yu Zhang, Qi Shi

Findings of ACL 2023 · Paper · Code

FAAST: Forward-Only Associative Learning via Closed-Form Fast Weights for Test-Time Supervised Adaptation

Guangsheng Bao, Hongbo Zhang, Han Cui, Ke Sun, Yanbin Zhao, Juncai He, Yue Zhang

arXiv preprint · Paper · Code

Minimizing Mismatch Risk: A Prototype-Based Routing Framework for Zero-shot LLM-generated Text Detection

Ke Sun, Guangsheng Bao, Han Cui, Yue Zhang

ICML 2026 · Paper

When AI Settles Down: Late-Stage Stability as a Signature of AI-Generated Text Detection

Ke Sun, Guangsheng Bao, Han Cui, Yue Zhang

arXiv preprint · Paper

Experience

Education

Harbin Institute of Technology, Harbin, Heilongjiang, China (2022.08 - 2024.12)

  • MS. in Computer Science and Technology, rank: 4/260
  • Research Center for Social Computing and Information Retrieval
  • Research Interests: Explicit Reasoning, Explanation Graph Generation, Natural Language Processing

Harbin Institute of Technology, Harbin, Heilongjiang, China (2018.08 - 2022.06)

  • BS. in Computer Science and Technology, Honors School

Internship

Westlake NLP Group, Westlake University (Research Intern)

  • March 2023 - Now · Hangzhou, Zhejiang, China

Foundation Research Center, PCG, Tencent (Algorithm Engineer Intern)

  • June 2023 - September 2023 · Shenzhen, Guangdong, China

Research Center for Social Computing and Information Retrieval (SCIR), HIT (Research Intern)

  • June 2021 - July 2021 · Harbin, Heilongjiang, China