Researcher — paper.dinq.me

2026 顶会论文

作者位次 #3/11ICLR 2026Poster· Reinforcement Learning / Everything Else

CDE: Curiosity-Driven Exploration for Efficient Reinforcement Learning in Large Language Models

Runpeng Dai · Linfeng Song · Haolin Liu · Zhenwen Liang · Dian Yu · Haitao Mi · Zhaopeng Tu · Rui Liu · Tong Zheng · Hongtu Zhu · Dong Yu

强化学习与后训练效率与系统OpenReview

一作ICLR 2026Poster· Theory / Reinforcement Learning and Planning

An Improved Model-free Decision-estimation Coefficient with Applications in Adversarial MDPs

Haolin Liu · Chen-Yu Wei · Julian Zimmert

推理安全强化学习与后训练OpenReview

作者位次 #4/9CVPR 2026Poster· Computer Vision & Pattern Recognition

Stable and Efficient Single-Rollout RL for Multimodal Reasoning

Rui Liu · Dian Yu · Lei Ke · Haolin Liu · Yujun Zhou · Zhenwen Liang · Haitao Mi · Pratap Tokekar · Dong Yu

多模态推理强化学习与后训练OpenReview 项目页

一作ACL 2026Findings· Findings of ACL

Save the Good Prefix: Precise Error Penalization via Process-Supervised RL to Enhance LLM Reasoning

Haolin Liu · Dian Yu · Sidi Lu · Yujun Zhou · Rui Liu · Zhenwen Liang · Haitao Mi · Chen-Yu Wei · Dong Yu

摘要

Reinforcement learning (RL) has emerged as a powerful framework for improving the reasoning capabilities of large language models (LLMs). However, most existing RL approaches rely on sparse outcome rewards, which fail to credit correct intermediate steps in partially successful solutions. Process...

强化学习与后训练对齐与偏好评测与基准OpenReview 项目页

Haolin Liu

获取完整研究档案

联系方式

2026 顶会论文

CDE: Curiosity-Driven Exploration for Efficient Reinforcement Learning in Large Language Models

An Improved Model-free Decision-estimation Coefficient with Applications in Adversarial MDPs

Stable and Efficient Single-Rollout RL for Multimodal Reasoning

Save the Good Prefix: Precise Error Penalization via Process-Supervised RL to Enhance LLM Reasoning

常见合作者