2 papers
cs.LG2023
Ternary Singular Value Decomposition as a Better Parameterized Form in Linear Mapping
Boyu Chen, Hanxuan Chen, Jiao He +2
We present a simple yet novel parameterized form of linear mapping to achieves remarkable network compression performance: a pseudo SVD called Ternary SVD (TSVD). Unlike vanilla SV…
stat.ML2023
Modify Training Directions in Function Space to Reduce Generalization Error
Yi Yu, Wenlian Lu, Boyu Chen
We propose theoretical analyses of a modified natural gradient descent method in the neural network function space based on the eigendecompositions of neural tangent kernel and Fis…