Researcher — paper.dinq.me

2026 顶会论文

作者位次 #5/5ICML 2026Spotlight· Deep Learning / Large Language Models

NorMuon: Making Muon more efficient and scalable

Zichong Li · Liming Liu · Chen Liang · Weizhu Chen · Tuo Zhao

效率与系统OpenReview

作者位次 #4/6ICML 2026Poster· Deep Learning / Large Language Models

Shuffle the Context: RoPE-Perturbed Self-Distillation for Long-Context Adaptation

Zichong Li · Chen Liang · Liliang Ren · Tuo Zhao · Yelong Shen · Weizhu Chen

强化学习与后训练长上下文与架构OpenReview

作者位次 #8/9ICML 2026Poster· Deep Learning / Large Language Models

Alternating Reinforcement Learning for Rubric-Based Reward Modeling in Non-Verifiable LLM Post-Training

Ran Xu · Tianci Liu · Zihan Dong · Tony Yu · Ilgee Hong · Carl Yang · Linjun Zhang · Tuo Zhao · Haoyu Wang

强化学习与后训练对齐与偏好OpenReview

作者位次 #5/6ICML 2026Poster· Deep Learning / Generative Models and Autoencoders

Generalization Bounds for Discrete Diffusion: Statistical Advantage of Masking

Zixuan Zhang · Hengyu Fu · Zhuoran Yang · Mengdi Wang · Tuo Zhao · Minshuo Chen

理论与学习动力学OpenReview

作者位次 #4/4ICML 2026Poster· Deep Learning / Large Language Models

Approximation of Log-Partition Function in Policy Mirror Descent Induces Implicit Regularization for LLM Post-Training

Zhenghao Xu · Qin Lu · Changlong Yu · Tuo Zhao

强化学习与后训练OpenReview

作者位次 #6/7ACL 2026Poster· Natural Language Processing

O pen R ubrics: Towards Scalable Synthetic Rubric Generation for Reward Modeling and LLM Alignment

Tianci Liu · Ran Xu · Tony Yu · Ilgee Hong · Carl Yang · Tuo Zhao · Haoyu Wang

摘要

Reward modeling lies at the core of reinforcement learning from human feedback (RLHF), yet most existing reward models rely on scalar or pairwise judgments that fail to capture the multifaceted nature of human preferences. Recent studies have explored rubrics-as-rewards (RaR) that uses structured...

对齐与偏好评测与基准社会与人本OpenReview 项目页

Tuo Zhao

获取完整研究档案

联系方式