Researcher — paper.dinq.me

2026 顶会论文

作者位次 #3/9ICML 2026Poster· General Machine Learning / Evaluation

Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression

Xiang Liu · Zhenheng Tang · Hong Chen · Peijie Dong · Zeyu Li · Xiuze Zhou · Bo Li · Xuming Hu · Xiaowen Chu

效率与系统评测与基准推理OpenReview

作者位次 #3/8CVPR 2026Spotlight· Computer Vision & Pattern Recognition

OralGPT-Plus: Learning to Use Visual Tools via Reinforcement Learning for Panoramic X-ray Analysis

Yuxuan Fan · Jing Hao · Hong Chen · Jiahao Bao · Yihua Shao · Yuci Liang · Kuo Feng Hung · Hao Tang

多模态强化学习与后训练OpenReview 项目页

作者位次 #3/11ACL 2026Findings· Findings of ACL

Thinking Economically: A Hierarchical Framework for Adaptive-Complexity Reasoning in LLM s

Yubo Gao · Haotian Wu · Hong Chen · Junquan Huang · Yibo Yan · Jungang Li · Zihao Dongfang · Sicheng Tao · PS Tan · Jie Zhang · Xuming Hu

摘要

Chain-of-Thought (CoT) has significantly enhanced LLM reasoning, yet often incurs substantial computational overhead due to “overthinking”: generating excessively long rationales without commensurate accuracy gains. Existing efficiency methods typically apply uniform compression, which overlooks...

推理效率与系统OpenReview 项目页

作者位次 #3/6AAAI 2026Poster· Artificial Intelligence

Deconstructing Pre-training: Knowledge Attribution Analysis in MoE and Dense Models

Bo Wang · Junzhuo Li · Hong Chen · Yuanlin Chu · Yuxuan Fan · Xuming Hu

效率与系统数据与预训练检索与知识OpenReview

Hong Chen

获取完整研究档案

联系方式

2026 顶会论文

Semantic Integrity Matters: Benchmarking and Preserving High-Density Reasoning in KV Cache Compression

OralGPT-Plus: Learning to Use Visual Tools via Reinforcement Learning for Panoramic X-ray Analysis

Thinking Economically: A Hierarchical Framework for Adaptive-Complexity Reasoning in LLM s

Deconstructing Pre-training: Knowledge Attribution Analysis in MoE and Dense Models

常见合作者