Han Liu

PhD student, Tsinghua University, Tsinghua University

Joined

April 2022

Names

Han Liu (Preferred)

Emails

****@mails.tsinghua.edu.cn (Confirmed)

Personal Links

ORCID

Semantic Scholar

Career & Education History

PhD student

Tsinghua University, Tsinghua University (tsinghua.edu.cn)

2018 – 2026

Intern

Beijing Baichuan Intelligence Technology Co., Ltd. (baichuan-ai.com)

2024 – 2025

Undergrad student

Tsinghua University, Tsinghua University (tsinghua.edu.cn)

2014 – 2018

Advisors, Relations & Conflicts

Coworker

Peng Cui

2024 – Present

Coworker

Bingning Wang

2020 – Present

PhD Advisor

Xiaolin Hu

2018 – Present

Expertise

Controllable Language Generation

2020 – Present

Image Captioning

2016 – 2019

Publications

Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers
Rihui Xin, Han Liu, Zecheng Wang, Yupeng Zhang, Dianbo Sui, Xiaolin Hu, Bingning Wang
- ICLR 2026 Workshop LLM Reasoning
- Readers: Everyone
Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers
Rihui Xin, Han Liu, Zecheng Wang, Yupeng Zhang, Dianbo Sui, Xiaolin Hu, Bingning Wang
- Submitted to ICLR 2026
- Readers: Everyone
VPO: Reasoning Preferences Optimization Based on $\mathcal{V}$-Usable Information
Zecheng Wang, Chunshan Li, Yupeng Zhang, Han Liu, Bingning Wang, Dianhui Chu, Dianbo Sui
- NeurIPS 2025 spotlight
- Readers: Everyone
Teaching Others is Teaching Yourself Regularization For Controllable Language Models
Han Liu, Bingning Wang, Ting Yao, Haijin Liang, Jianjin Xu, Xiaolin Hu
- Published: 01 Feb 2023, Last Modified: 13 Feb 2023
- Submitted to ICLR 2023
- Readers: Everyone
Vocabulary-Wide Credit Assignment for Training Image Captioning Models
Han Liu, Shifeng Zhang, Ke Lin, Jing Wen, Jianmin Li, Xiaolin Hu
- 2021 (modified: 03 Nov 2022)
- IEEE Trans. Image Process. 2021
- Readers: Everyone

Han Liu

Names

Emails

Personal Links

Career & Education History

Advisors, Relations & Conflicts

Expertise

Publications

Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers

Surrogate Signals from Format and Length: Reinforcement Learning for Solving Mathematical Problems without Ground Truth Answers

VPO: Reasoning Preferences Optimization Based on $\mathcal{V}$-Usable Information

Teaching Others is Teaching Yourself Regularization For Controllable Language Models

Vocabulary-Wide Credit Assignment for Training Image Captioning Models

Co-Authors