MIAO'S GROUP

Research Aim: (Less is More)

Currently, my research interests are especially shifting to Foundation Model Compression and Acceleration, from theory to algorithm and system. My current students are working on:

• Theoretical Understanding

Jiarui Jiang(Attention Machanism), Shiyuan Ren(Sparse training), Hongyi Wan(Continual Learning and Unlearning)

• Model Compression

Jiaqi Zhao(Quantization), Chao Zeng(Quantization), Ming Wang(Sparse Attention), Yixuan Dong(Quantization Kernel Design), Ruizi Han(Sparse Kernel Design)

• Continual Learning/Unlearning

Qianlong Xiang(Diffusion Unlearning), Haomiao Qiu(Continual Learning and Unlearning)

• Applications

Ao Jin(Long Video Understanding), Junpeng Jiang(Long Video Genenration), Lexiao Zou(GUI Agent), Lirong Jie(GUI Agent), Ji Shi(Code Agent), Liyang Zheng(TinyML), Yanda Chen(Dataset Distillation)