paper.dinq.meAI 与前沿科学顶会 · 学术新星雷达
登录 DINQ
返回新星榜
FH

Fei Huang

ICLR 2026×20AAAI 2026×2COLM 2026智能体与工具多模态强化学习与后训练检索与知识数据与预训练
OROpenReview
新星分资深
56
一作贡献0
核心/参与122
方向广度16
合作网络12
0 一作 · 23 篇 · ICLR 2026 / AAAI 2026 / COLM 2026

获取完整研究档案

首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。

联系方式

实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。

2026 顶会论文

23
作者位次 #6/19COLM 2026Poster· Poster Session 6· Franciscan C· #98

WorldPM: Scaling Human Preference Modeling via Real-World Feedback

Binghai Wang · Runji Lin · Keming Lu · Le Yu · Zhenru Zhang · Fei Huang · Chujie Zheng · Kai Dang · Yang Fan · An Yang · Binyuan Hui · Dayiheng Liu · Tao Gui · Qi Zhang · Xuanjing Huang · Yu-Gang Jiang · Bowen Yu · Jingren Zhou · Junyang Lin

We collect massive human preference data from forums to scale human preference modeling, and investigate the scaling patterns across several dimensions.

摘要

Inspired by the success of scaling language modeling through massive real-world corpora, we demonstrate that similar power-law patterns exist in preference modeling. We propose \textbf{World Preference Modeling (WorldPM)}, which model a robust consensus on intrinsic utility from real-world user interactions. We curate a 15M-scale preference dataset from online forum data and conduct large-scale experiments on models from 1.5B to 72B parameters. Across 10 objective preference test sets with consensus-based labels, we observe two distinct scaling patterns: (1) common-sense preference (e.g., identifying factual errors) exhibits stable log-linear improvement across all model sizes, while (2) domain preference (e.g., coding, math, knowledge QA) reveals an emergent phenomenon where only sufficiently large models achieve meaningful scaling. Through training dynamics analysis, we reveal the learning mechanism behind preference scaling: the model first learns to discriminate using surface preferences, but counter-surface examples progressively compel it to discover intrinsic utility beyond surface features as training scales. Further experiments validate WorldPM as an effective foundation for preference fine-tuning, broadly improving generalization across human preference datasets of varying sizes (7K, 100K, and 800K), with gains exceeding 5\% on many key subtasks.

对齐与偏好数据与预训练检索与知识reward modeling; human preference; scaling law; RLHF; AlignmentOpenReview
作者位次 #17/21ICLR 2026Poster

WebSailor-V2: Bridging the Chasm to Proprietary Agents via Synthetic Data and Scalable Reinforcement Learning

Kuan Li · Zhongwang Zhang · Huifeng Yin · Rui Ye · Yida Zhao · Liwen Zhang · Litu Ou · Ding-Chu Zhang · Xixi Wu · Xinmiao Yu · Jialong Wu · Xinyu Wang · Zile Qiao · Zhen Zhang · Yong Jiang · Pengjun Xie · Fei Huang · Zhiqin Xu · Shuai Wang · Minhao Cheng · Jingren Zhou

智能体与工具强化学习与后训练数据与预训练OpenReview
作者位次 #2/3AAAI 2026Poster· Artificial Intelligence

Selective Weak-to-Strong Generalization

Hao Lang · Fei Huang · Yongbin Li

理论与学习动力学OpenReview

常见合作者

JZJingren Zhou×10YJYong Jiang×10PXPengjun Xie×10XWXinyu Wang×7KLKuan Li×6MYMing Yan×6YLYongbin Li×5ZQZile Qiao×5HYHuifeng Yin×5ZZZhen Zhang×4YZYida Zhao×4LZLiwen Zhang×4