I am a second-year PhD student at Xidian University, advised by Professor Cheng Deng. My research focuses on Multimodal Large Language Models (MLLM) and Human Motion Generation and Understanding.
🔥 News
- 2026.08 Released gestureopd.
- 2026.08 ✅ Paper accepted by ACMMM 2026.
- 2026.04 ✅ Paper accepted by ACL Findings 2026.
- 2025.04 ✅ Paper accepted by ICLR 2025.
- 2023.06 🎓 Obtained Bachelor’s degree from Wuhan University of Technology.
- 2023.05 ✅ Paper accepted by PR 2023.
- 2022.03 ✅ Paper accepted by ICME 2022.
📝 Publications
Multimodal Large Language Models (MLLM) & Hallucination
-
Revealing and Enhancing Core Visual Regions: Harnessing Internal Attention Dynamics for Hallucination Mitigation in LVLMs
Guangtao Lyu, Qi Liu, Cheng-Zhong Xu, Jiexi Yan, Muli Yang, Xueting Li, Fen Fang, Cheng Deng
ACL Findings 2026 [PDF] -
Semantically Comprehensive Token Pruning in LVLMs via Maximizing Concept Coverage
Xueting Li, Qi Liu, Chenghao Xu, Xu Yang, Guangtao Lyu, Jiahua Li, Cheng Deng
ACL 2026 [PDF] -
Fisher-Driven Adaptive Locating for Knowledge Editing in Large Language Models
Chenghao Xu, Jiexi Yan, Guangtao Lyu, Qi Liu, Muli Yang, Cheng Deng
ACL 2026 [PDF] -
Channel-masked Asymmetric Distribution Matching for Cross-Domain Generalized Dataset Distillation
Qi Liu, Chenghao Xu, Jiexi Yan, Guangtao Lyu, Erkun Yang, Guihai Chen, Yanhua Yang
AAAI 2026 [PDF] -
Towards Interpretable Hallucination Analysis and Mitigation in LVLMs via Contrastive Neuron Steering
Guangtao Lyu, Xinyi Cheng, Qi Liu, Chenghao Xu, Jiexi Yan, Muli Yang, Fen Fang, Cheng Deng
arXiv 2026 [PDF] -
Revealing Perception and Generation Dynamics in LVLMs: Mitigating Hallucinations via Validated Dominance Correction
Guangtao Lyu, Xinyi Cheng, Chenghao Xu, Qi Liu, Muli Yang, Fen Fang, Huilin Chen, Jiexi Yan, Xu Yang, Cheng Deng
arXiv 2025 [PDF]
Human Motion Generation and Understanding
-
Towards Unified Human Motion-Language Understanding via Sparse Interpretable Characterization
Guangtao Lyu, Chenghao Xu, Jiexi Yan, Muli Yang, Cheng Deng
ICLR 2025 [PDF] [Project] -
COME: Advancing Motion Representation and Generative Modeling for High-Quality Text-to-Motion Generation
Guangtao Lyu, Xinyi Cheng, Qi Liu, Chenghao Xu, Jiexi Yan, Muli Yang, Fen Fang, Cheng Deng
ACMMM 2026 -
Smooth and Flexible Camera Movement Synthesis via Temporal Masked Generative Modeling
Chenghao Xu, Guangtao Lyu, Jiexi Yan, Muli Yang, Cheng Deng
NeurIPS 2025 [PDF] -
LLM Knows Body Language, Too: Translating Speech Voices into Human Gestures
Chenghao Xu, Guangtao Lyu, Jiexi Yan, Muli Yang, Cheng Deng
ACL 2024 [PDF] -
Beyond Global Alignment: Fine-Grained Motion-Language Retrieval via Pyramidal Shapley-Taylor Learning
Hanmo Chen, Guangtao Lyu, Chenghao Xu, Jiexi Yan, Xu Yang, Cheng Deng
ICML 2026 [PDF] -
Tempo as the Stable Cue: Hierarchical Mixture of Tempo and Beat Experts for Music to 3D Dance Generation
Guangtao Lyu, Chenghao Xu, Qi Liu, Jiexi Yan, Muli Yang, Fen Fang, Cheng Deng
arXiv 2025 [PDF] -
Towards Arbitrary Motion Completing via Hierarchical Continuous Representation
Chenghao Xu, Guangtao Lyu, Qi Liu, Jiexi Yan, Muli Yang, Cheng Deng
arXiv 2025 [PDF]
Scene Text Removal
-
FETNet: Feature Erasing and Transferring Network for Scene Text Removal
Guangtao Lyu, Kun Liu, Anna Zhu, Seiichi Uchida, Brian Kenji Iwana
Pattern Recognition 2023 [PDF] [Project] -
PSSTRNet: Progressive Segmentation-Guided Scene Text Removal Network
Guangtao Lyu, Anna Zhu
ICME 2022 [PDF] [Code] -
HelixNet: Dual Helix Cooperative Decoders for Scene Text Removal
Kun Liu, Guangtao Lyu, Anna Zhu
PRCV 2023 [PDF] -
MSLKANet: A Multi-Scale Large Kernel Attention Network for Scene Text Removal
Guangtao Lyu
arXiv 2022 [PDF]
🎖 Honors and Awards
- 2023.9 2023IKCEST “一带一路”国际大数据竞赛 (Rank 4th) 链接（¥1万）
- 2023.9 2023-finvcup 信也科技杯 (Rank 2nd) 链接（¥8万）
- 2023.9 2023-Video semantic understanding 视频语义理解 (Rank 2nd) 链接（¥1.5万）
- 2022.8 WeChat-Big-Data-Challenge-2022 微信大数据挑战赛 (Rank 3rd) 链接（该届冠军为郭达雅，即 DeepSeek-R1 作者）（¥6万）
📖 Educations
- 2023.09 - present, PhD, Xidian University, Xi’an.
- 2019.09 - 2023.06, Undergraduate, Wuhan University of Technology, Wuhan.
💻 Internships
- None