NSFC-12571561 (General Program): Compression of Large-scale Transformer-based Models with Random Matrix Methods
PI, National Natural Science Foundation of China General Program, Compression of Large-scale Transformer-based Models with Random Matrix Methods: From Scaling Law to Dynamic Adaptive Model Compression (NSFC-12571561), CNY 440K, with Cosme Louart as co-PI.
National Key R&D Program of China: Foundations of Machine Learning for Understanding Large Models
Co-PI, National Key R&D Program of China (Youth Scientist Project), Foundations of Machine Learning for Understanding Large Models (2025YFA1018600), CNY 1M/3M. PI:Jun Shu, Xi’an Jiaotong University.
HUST Basic Research Support Program: Theory and Methods for Large-scale Machine Learning via RMT
PI, HUST Basic Research Support Program (Mathematics Track), Theory and Methods for Large-scale Machine Learning via RMT (2025BRSXB0004), CNY 380K.
Completed
Guangdong Key Lab Open Fund OFA00003: Generalization Theory for Transformer-based Models
PI, Guangdong Key Lab of Mathematical Foundations for Artificial Intelligence Open Fund, Generalization Theory for Transformer-based Models via Random Matrix Methods (OFA00003), CNY 100K, with Jeff Yao as co-PI.
NSFC-62206101 (Youth Program): Fundamental Limits of Pruning Deep Neural Network Models
PI, National Natural Science Foundation of China Youth Program, Fundamental Limits of Pruning Deep Neural Network Models via Random Matrix Methods (NSFC-62206101), CNY 300K.
CCF-Hikvision Open Fund 20210008: Random Matrix Theory and Information Bottleneck
PI, CCF-Hikvision Open Fund, Random Matrix Theory and Information Bottleneck for Neural Network Compression (20210008), CNY 280K, with Kai Wan as co-PI.