李晓鹏｜武汉大学｜学者主页

2026

基于自适应大核卷积和特征融合的视线估计

李晓鹏，王演，刘教

《计算机工程与设计》

DOI

2026

AFFLIE: Adaptive feature fusion for low-light image enhancement

Yaxin Lin, Xiaopeng Li, Lian Zou, Liqing Zhou, Cien Fan

Displays

DOI

2026

SPAR: Step-wise Path Dispatching and Asymmetric Re-routing for Efficient MoE Inference

Qing-Xiao Zhang, Xiaopeng Li, Jin-Zhu Kong, Xiao-Dong Liu, Bing Ji, Shasha Li, Jun Ma, Jie Yu

International Conference on Intelligent Computing

DOI

2026

Leveraging Image as Compressed Visual Prompt and Hierarchical Visual Knowledge for Effective Image Utilization in MLLMs

She-Zheng Song, Kangcheng Ding, Shan Zhao, Shasha Li, Xiaopeng Li, Cheng-Yu Wang, Qian Wan, Bing Ji, Jie Yu

AAAI Conference on Artificial Intelligence

DOI

2026

JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path Rectification

Xi Wang, Song-Lei Jian, Shasha Li, Xiaopeng Li, Zhao-Ye Li, Bing Ji, Bao-Sheng Wang, Jie Yu

Annual Meeting of the Association for Computational Linguistics

DOI

2026

SEAttack: A self-evolving jailbreak attack to induce toxic responses for non-toxic queries in large language models

Huijun Liu, Shasha Li, Bin Ji, Xiao-Hu Du, Xiaopeng Li, Jun Ma, Jie Yu

Information Processing & Management

DOI

2026

SkillSight: Calibrating Generic Content Bias for Skill Retrieval

Jin-Ying Xiao, Bin Li, Xiaopeng Li, Jian-Ling Li, Jiacheng Jie, Xiao-Dong Liu, Ma Jun, Chao Wang, Nyima Tashi, Jie Yu

arXiv.org

DOI

2026

Where Does Vision Meet Language? Understanding and Refining Visual Fusion in MLLMs via Contrastive Attention

She-Zheng Song, Shasha Li, Shan Zhao, Xiaopeng Li, Qian Wan, Cheng-Yu Wang, Tian-Wei Yan, Jun Ma, Jie Yu

arXiv.org

DOI

2026

Efficient Exploration for Iterative Nash Preference Optimization

Tian-Long Nan, Xiaopeng Li, C. Kroer, Tian-Yi Lin

arXiv.org

DOI

2026

Reward-free Alignment for Conflicting Objectives

Peter Chen, Xiaopeng Li, Xi Chen, Tian-Yi Lin

arXiv.org

DOI

2026

KnowFC: Navigating Knowledge Conflicts in Large Language Model-based Fact-Checking

Yue Zhang, Shicheng Zhou, Xiaopeng Li, Zhiliang Tian, Yifu Gao, Shi-Yi Zhang, Wenqing Hou, Yuying Liu, Bin Zhou

Web Search and Data Mining

DOI

2026

When Hard Negatives Hurt: Bridging the Generative Discriminative Gap in Hard Negative Synthesis for Retrieval

Zhi-Cheng Zhang, Jiwei Tang, Kui-Cai Dong, Xiaopeng Li, Jieming Zhu, Jingyu Li, Qianhui Zhu, Fengyu Lu, Jia-Heng Wang, Gang Wang, Hai-Tao Zheng, Zhao-Cheng Du

Proceedings of the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining V.2

DOI

2025

基于深度学习的鲁棒无载体图像隐写方法研究

李晓鹏

《兰州理工大学》

2025

Rethinking Residual Distribution in Locate-then-Edit Model Editing

Xiaopeng Li, Shangwen Wang, Shasha Li, She-Zheng Song, Bing Ji, Ma Jun, Jie Yu

Neural Information Processing Systems

DOI

2025

How to Bridge the Gap Between Modalities: Survey on Multimodal Large Language Model

She-Zheng Song, Xiaopeng Li, Sha-Sha Li, Shan Zhao, Jie Yu, Jun Ma, Xiaoguang Mao, Weimin Zhang, Meng Wang

IEEE Transactions on Knowledge and Data Engineering

DOI

2025

Stand on The Shoulders of Giants: Building JailExpert from Previous Attack Experience

Xi Wang, Songlei Jian, Shasha Li, Xiaopeng Li, Bing Ji, Ma Jun, Xiaodong Liu, Jing Wang, Jian-Feng Zhang, Jie Yu, Feilong Bao, Wangbaosheng

Conference on Empirical Methods in Natural Language Processing

DOI

2025

SWSC: Compressing Large Language Models via Sharing Weights for Similar Channels

Binrui Zeng, Yong-Tao Tang, Bing Ji, Xiaodong Liu, Xiaopeng Li, Jie Yu

International Conference on Computer Science and Artificial Intelligence

DOI

2025

EMSEdit: Efficient Multi-Step Meta-Learning-based Model Editing

Xiaopeng Li, Shasha Li, Xi Wang, She-Zheng Song, Bing Ji, Shangwen Wang, Jun Ma, Xiao-Dong Liu, Mina Liu, Jie Yu

The Web Conference

DOI

2025

MapAgent: Trajectory-Constructed Memory-Augmented Planning for Mobile Task Automation

Yi-Bo Kong, Dian-xi Shi, Guo-Li Yang, Kedi Zhang, Chen Huang, Xiaopeng Li, Song-Chang Jin

arXiv.org

DOI

2025

SWSC: Shared Weight for Similar Channel in LLM

Binrui Zeng, Yong-Tao Tang, Xiaodong Liu, Xiaopeng Li

arXiv.org

DOI

2025

Stepwise Guided Policy Optimization: Coloring Your Incorrect Reasoning in GRPO

Peter Chen, Xiaopeng Li, Zi-Niu Li, Xi Chen, Tian-Yi Lin

Trans. Mach. Learn. Res.

2025

Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward

Peter Chen, Xiaopeng Li, Zi-Niu Li, Wo-Tao Yin, Xi Chen, Tian-Yi Lin

arXiv.org

DOI

2024

Model Editing for LLMs4Code: How Far are we?

Xiaopeng Li, Shangwen Wang, Shasha Li, Jun Ma, Jie Yu, Xiaodong Liu, Jing Wang, Bing Ji, Weimin Zhang

International Conference on Software Engineering

DOI

2024

SWEA: Updating Factual Knowledge in Large Language Models via Subject Word Embedding Altering

Xiaopeng Li, Shasha Li, She-Zheng Song, Huijun Liu, Bing Ji, Xi Wang, Jun Ma, Jie Yu, Xiao-Dong Liu, Jing Wang, Weimin Zhang

AAAI Conference on Artificial Intelligence

DOI

2024

PTA: Enhancing Multimodal Sentiment Analysis through Pipelined Prediction and Translation-based Alignment

She-Zheng Song, Shasha Li, Shan Zhao, Cheng-Yu Wang, Xiaopeng Li, Jie Yu, Qian Wan, Jun Ma, Tian-Wei Yan, Wen-Tao Ma, Xiaoguang Mao

arXiv.org

DOI

2024

DWE+: Dual-Way Matching Enhanced Framework for Multimodal Entity Linking

She-Zheng Song, Shasha Li, Shan Zhao, Xiaopeng Li, Cheng-Yu Wang, Jie Yu, Jun Ma, Tian-Wei Yan, Bing Ji, Xiaoguang Mao

arXiv.org

DOI

2024

DIM: Dynamic Integration of Multimodal Entity Linking with Large Language Model

She-Zheng Song, Shasha Li, Jie Yu, Shan Zhao, Xiaopeng Li, Jun Ma, Xiao-Dong Liu, Zhuo Li, Xiaoguang Mao

Chinese Conference on Pattern Recognition and Computer Vision

DOI

2024

Identifying Knowledge Editing Types in Large Language Models

Xiaopeng Li, Shasha Li, Shangwen Wang, She-Zheng Song, Bing Ji, Huijun Liu, Jun Ma, Jie Yu

Knowledge Discovery and Data Mining

DOI

2024

MOSABench: Multi-Object Sentiment Analysis Benchmark for Evaluating Multimodal Large Language Models Understanding of Complex Image

She-Zheng Song, Cheng-Xiang He, Shasha Li, Shan Zhao, Cheng-Yu Wang, Tian-Wei Yan, Xiaopeng Li, Qian Wan, Jun Ma, Jie Yu, Xiaoguang Mao

arXiv.org

DOI

2024

LSAQ: Layer-Specific Adaptive Quantization for Large Language Model Deployment

Binrui Zeng, Bing Ji, Xiao-Dong Liu, Jie Yu, Shasha Li, Jun Ma, Xiaopeng Li, Shangwen Wang, Xinran Hong, Yong-Tao Tang

IEEE International Joint Conference on Neural Network

DOI

2024

Socratic Human Feedback (SoHF): Expert Steering Strategies for LLM Code Generation

Subramanian Chidambaram, Er-Ran L. Li, Min Bai, Xiaopeng Li, Kaixiang Lin, Xiong Zhou, Alex C. Williams

Conference on Empirical Methods in Natural Language Processing

DOI

2023

PMET: Precise Model Editing in a Transformer

Xiaopeng Li, Shasha Li, She-Zheng Song, Jing Yang, Jun Ma, Jie Yu

AAAI Conference on Artificial Intelligence

DOI

李晓鹏

精选作品

期刊论文

基于自适应大核卷积和特征融合的视线估计

AFFLIE: Adaptive feature fusion for low-light image enhancement

SPAR: Step-wise Path Dispatching and Asymmetric Re-routing for Efficient MoE Inference

Leveraging Image as Compressed Visual Prompt and Hierarchical Visual Knowledge for Effective Image Utilization in MLLMs

JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path Rectification

SEAttack: A self-evolving jailbreak attack to induce toxic responses for non-toxic queries in large language models

SkillSight: Calibrating Generic Content Bias for Skill Retrieval

Where Does Vision Meet Language? Understanding and Refining Visual Fusion in MLLMs via Contrastive Attention

Efficient Exploration for Iterative Nash Preference Optimization

Reward-free Alignment for Conflicting Objectives

KnowFC: Navigating Knowledge Conflicts in Large Language Model-based Fact-Checking

When Hard Negatives Hurt: Bridging the Generative Discriminative Gap in Hard Negative Synthesis for Retrieval

基于深度学习的鲁棒无载体图像隐写方法研究

Rethinking Residual Distribution in Locate-then-Edit Model Editing

How to Bridge the Gap Between Modalities: Survey on Multimodal Large Language Model

Stand on The Shoulders of Giants: Building JailExpert from Previous Attack Experience

SWSC: Compressing Large Language Models via Sharing Weights for Similar Channels

EMSEdit: Efficient Multi-Step Meta-Learning-based Model Editing

MapAgent: Trajectory-Constructed Memory-Augmented Planning for Mobile Task Automation

SWSC: Shared Weight for Similar Channel in LLM

Stepwise Guided Policy Optimization: Coloring Your Incorrect Reasoning in GRPO

Exploration vs Exploitation: Rethinking RLVR through Clipping, Entropy, and Spurious Reward

Model Editing for LLMs4Code: How Far are we?

SWEA: Updating Factual Knowledge in Large Language Models via Subject Word Embedding Altering

PTA: Enhancing Multimodal Sentiment Analysis through Pipelined Prediction and Translation-based Alignment

DWE+: Dual-Way Matching Enhanced Framework for Multimodal Entity Linking

DIM: Dynamic Integration of Multimodal Entity Linking with Large Language Model

Identifying Knowledge Editing Types in Large Language Models

MOSABench: Multi-Object Sentiment Analysis Benchmark for Evaluating Multimodal Large Language Models Understanding of Complex Image

LSAQ: Layer-Specific Adaptive Quantization for Large Language Model Deployment

Socratic Human Feedback (SoHF): Expert Steering Strategies for LLM Code Generation

PMET: Precise Model Editing in a Transformer

研究方向