Researcher — paper.dinq.me

2026 顶会论文

作者位次 #2/6ICML 2026Poster· Deep Learning / Large Language Models

NITP: Next Implicit Token Prediction for LLM Pre-training

Xiangdong Zhang · Debing Zhang · Shaofeng Zhang · Xiaohan Qin · Yu Cheng · Junchi Yan

数据与预训练OpenReview

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

PretrainZero: Reinforcement Active Learning on Pretraining Data

Xingrun Xing · Zhiyuan Fan · Jie Lou · Guoqi Li · Jiajun Zhang · Debing Zhang

强化学习与后训练数据与预训练OpenReview

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

Coupled Variational Reinforcement Learning for Language Model General Reasoning

Xueru Wen · Jie Lou · Yanjiang Liu · Hongyu Lin · Ben He · Xianpei Han · Le Sun · Yaojie Lu · Debing Zhang

推理强化学习与后训练OpenReview

作者位次 #8/11ICML 2026Poster· Deep Learning / Large Language Models

Tackling Length Inflation Without Trade-offs: Group Relative Reward Rescaling for Reinforcement Learning

Zichao Li · Jie Lou · Fangchen Dong · Zhiyuan Fan · Mengjie Ren · Hongyu Lin · Xianpei Han · Debing Zhang · Le Sun · Yaojie Lu · XingYu

强化学习与后训练OpenReview

作者位次 #7/8ICML 2026Poster· Deep Learning / Large Language Models

Inner-layer Token Self-modulation as Another Scaling Axis for LLMs

Yebin Yang · Huaijin Wu · Jingtao Han · Yu Wang · Xiaohan Qin · Jingzhi Wang · Debing Zhang · Junchi Yan

OpenReview

作者位次 #8/9ACL 2026Findings· Findings of ACL

M em S earcher: Iterative Memory Integration for Search Agent via End-to-End Reinforcement Learning

Qianhao Yuan · Jie Lou · Zichao Li · Jiawei Chen · Yaojie Lu · Hongyu Lin · Le Sun · Debing Zhang · Xianpei Han

摘要

LLM-based search agents often concatenate the full interaction history into the context, producing long and noisy inputs, and increasing compute cost and GPU memory overhead. To address this issue, we propose MemSearcher, an agent framework that maintains a compact memory during multi-turn intera...

强化学习与后训练检索与知识智能体与工具OpenReview 项目页

Debing Zhang

获取完整研究档案

联系方式