CVPR 2025: 'Unified Dense Prediction of Video Diffusion'
CVPR 2024 (*equal contribution): 'UniGS: Unified Representation for Image Generation and Segmentation'
NeurIPS 2025: 'Distort Time to Improve Video Temporal Reasoning'
ArXiv (*equal contribution): 'Generalizable entity grounding via assistance of large language model'
Under review: 'VRMDiff: Text-Guided Video Referring Matting Generation of Diffusion'
Served as reviewer for top-tier conferences (CVPR, ECCV, NeurIPS, ICLR, ICML, etc.) and journals (TPAMI, Neurocomputing); Top Reviewer at NeurIPS'24,25