Pengyu Cheng

Research, Alibaba Group

Joined

July 2019

Names

Pengyu Cheng

Emails

****@duke.edu (Confirmed)

****@tencent.com (Confirmed)

****@gmail.com (Confirmed)

Personal Links

Career & Education History

Research

Alibaba Group (alibaba-inc.com)

2025 – Present

Researcher

Tencent (tencent.com)

2021 – 2024

PhD student

Duke University (duke.edu)

2017 – 2021

Advisors, Relations & Conflicts

PhD Advisor

Lawrence Carin

2017 – 2021

Expertise

Reinforcement Learning

2023 – Present

Natural Lauguage Processing

2018 – Present

Deep Learning

2015 – Present

Publications

Adversarial Preference Optimization: Enhancing Your Alignment via RM-LLM Game
Pengyu Cheng, Yifan Yang, Jian Li, Yong Dai, Tianhao Hu, Peixin Cao, nan du, Xiaolong Li
- OpenReview Archive Direct Upload
- Readers: Everyone
Search Self-Play: Pushing the Frontier of Agent Capability without Supervision
Hongliang Lu, Yuhang Wen, Pengyu Cheng, Ruijin Ding, Jiaqi Guo, Haotian Xu, Chutian Wang, Haonan Chen, xiaoxi jiang, guanjunjiang
- ICLR 2026 Poster
- Readers: Everyone
Eliminating Inductive Bias in Reward Models with Information-Theoretic Guidance
Zhuo Li, Pengyu Cheng, Zhechao Yu, FeifeiTong, Anningzhe Gao, Tsung-Hui Chang, Xiang Wan, erchao.zec, xiaoxi jiang, guanjunjiang
- ICLR 2026 Poster
- Readers: Everyone
RLHF in an SFT Way: From Optimal Solution to Reward-Weighted Alignment
Yuhao Du, Zhuo Li, Pengyu Cheng, Zhihong Chen, Yuejiao XIE, Xiang Wan, Anningzhe Gao
- Accepted by TMLR
- Readers: Everyone
Self-playing Adversarial Language Game Enhances LLM Reasoning
Pengyu Cheng, Tianhao Hu, Han Xu, Zhisong Zhang, Yong Dai, Lei Han, nan du, Xiaolong Li
- NeurIPS 2024 poster
- Readers: Everyone
Everyone Deserves A Reward: Learning Customized Human Preferences
Pengyu Cheng, Jiawen Xie, Ke Bai, Yong Dai, nan du
- Submitted to ICLR 2024
- Readers: Everyone
Chunk, Align, Select: A Simple Long-sequence Processing Method for Transformers
Jiawen Xie, Pengyu Cheng, Xiao Liang, Yong Dai, nan du
- ICLR 2024 Conference Withdrawn Submission
- Readers: Everyone
Toward Fairness in Text Generation via Mutual Information Minimizationbased on Importance Sampling
Rui Wang, Pengyu Cheng, Ricardo Henao
- 12 May 2023
- OpenReview Archive Direct Upload
- Readers: Everyone
Estimating Total Correlation with Mutual Information Bounds
Pengyu Cheng, Weituo Hao, Lawrence Carin
- Published: 07 Nov 2020, Last Modified: 05 May 2023
- NeurIPSW 2020: DL-IG Poster
- Readers: Everyone
FairFil: Contrastive Neural Debiasing Method for Pretrained Text Encoders
Pengyu Cheng, Weituo Hao, Siyang Yuan, Shijing Si, Lawrence Carin
- Published: 12 Jan 2021, Last Modified: 03 Apr 2024
- ICLR 2021 Poster
- Readers: Everyone

View all 25 publications

Co-Authors

View all 66 co-authors