Researcher — paper.dinq.me

2026 顶会论文

作者位次 #4/5NeurIPS 2026Poster

When and Why Does Multi-Agent Debate Fail and Does It Really Underperform?

Yongqiang Chen · Gang Niu · James Cheng · Bo Han · Masashi Sugiyama

智能体与工具OpenReview

作者位次 #10/12NeurIPS 2026Poster· Deep Learning / Algorithms

TriSpec: Ternary Speculative Decoding via Lightweight Proxy Verification

Haoyun Jiang · Junqi He · Feng Hong · Xinlong Yang · jianwei zhang · Zhengyang Zhuge · Zheng Li · Xiaofeng Cao · Zhiyong Chen · Bo Han · Junyang Lin · Jiangchao Yao

效率与系统代码OpenReview

作者位次 #5/5NeurIPS 2026Poster· Applications / Language, Speech and Dialog

Rethinking How to Remember: Beyond Atomic Facts in Lifelong LLM Agent Memory

Jingwei Sun · Jianing Zhu · Jiangchao Yao · Tongliang Liu · Bo Han

检索与知识智能体与工具多模态OpenReview

作者位次 #2/8NeurIPS 2026Poster

RewardFlow: Topology-Aware Reward Propagation on State Graphs for Agentic RL with LLMs

Xiao Feng · Bo Han · Zhanke Zhou · Jiaqi Fan · Jiangchao Yao · Ka H Li · Dahai Yu · Michael Ng

智能体与工具强化学习与后训练OpenReview

作者位次 #3/6NeurIPS 2026Poster· Applications / Chemistry and Drug Discovery

Adaptive Multi-Frame Learning for Expressive and Stable Atomic Representations

Jun Wang · Yifan Zeng · Bo Han · Fengwang Li · Aoni Xu · Tongliang Liu

可解释性理论与学习动力学OpenReview

作者位次 #4/5NeurIPS 2026Poster

UnlearningSoup: Is Repeated Tuning Necessary for Large Language Model Unlearning?

Puning Yang · Qizhou Wang · Junchi Yu · Bo Han · Xiuying Chen

安全OpenReview

作者位次 #4/5NeurIPS 2026Poster

Visual Harness: Grounding Multimodal Reasoning in Physics Engines

Chentao Cao · Zhanke Zhou · Sangni Duan · Bo Han · Hang Li

多模态推理OpenReview

作者位次 #5/6NeurIPS 2026Poster

Bayes-Sufficient Compression Is Not Enough: How Communication Helps in Multi-Agent Systems?

Yi Xie · Zhanke Zhou · Yi Fan · Yong Ge · Bo Han · Bo Liu

智能体与工具效率与系统OpenReview

作者位次 #5/6NeurIPS 2026Poster

Rethinking Parallel Multi-Agent Systems: A Cost-Aware Framework for Efficient Coordination

Yexiong Lin · Shanshan Ye · Yu Yao · Zhen Fang · Bo Han · Tongliang Liu

智能体与工具效率与系统OpenReview

作者位次 #6/6NeurIPS 2026Poster

Not All Low-Confidence Tokens Are Equal: Calibrated Confidence for Efficient Test-Time Reasoning

Tangyu Jiang · Haodi Wang · Yuanbing Zhu · Xiaojiang Du · Xiaohua Jia · Bo Han

推理效率与系统OpenReview

作者位次 #6/6NeurIPS 2026Poster

Transition-Aware Credit Assignment in Agentic Learning for LLM Reasoning

Xiangyu Lu · Zhanke Zhou · Jiazhe Ning · Chentao Cao · Jiangchao Yao · Bo Han

推理智能体与工具OpenReview

作者位次 #6/6NeurIPS 2026Poster

On the Information Loss of Multi-Token Prediction: Origin and Solution

Tangyu Jiang · Zhanke Zhou · Haodi Wang · Yiu-ming Cheung · Xiaohua Jia · Bo Han

OpenReview

作者位次 #7/7NeurIPS 2026Poster· Social Aspects / Accountability, Transparency and Interpretability

Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation-Level Enforcement

Xutao Mao · Jianing Zhu · Jinman Zhao · Tongliang Liu · Xiaowen Chu · Cong Wang (Key Laboratory of Computing Power Network and Information Security, Ministry of Education, Shandong Computer Science Center (National Supercomputer Center in Jinan), Qilu University of Technology (Shandong Academy of Sciences) · Bo Han

可解释性推理多模态OpenReview

Bo Han

获取完整研究档案

联系方式

2026 顶会论文

When and Why Does Multi-Agent Debate Fail and Does It Really Underperform?

TriSpec: Ternary Speculative Decoding via Lightweight Proxy Verification

Rethinking How to Remember: Beyond Atomic Facts in Lifelong LLM Agent Memory

RewardFlow: Topology-Aware Reward Propagation on State Graphs for Agentic RL with LLMs

Adaptive Multi-Frame Learning for Expressive and Stable Atomic Representations

UnlearningSoup: Is Repeated Tuning Necessary for Large Language Model Unlearning?

Visual Harness: Grounding Multimodal Reasoning in Physics Engines

Bayes-Sufficient Compression Is Not Enough: How Communication Helps in Multi-Agent Systems?

Rethinking Parallel Multi-Agent Systems: A Cost-Aware Framework for Efficient Coordination

Not All Low-Confidence Tokens Are Equal: Calibrated Confidence for Efficient Test-Time Reasoning

Transition-Aware Credit Assignment in Agentic Learning for LLM Reasoning

On the Information Loss of Multi-Token Prediction: Origin and Solution

Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation-Level Enforcement

常见合作者