Concept迹范数 / 核范数 / trace norm

Trace Norm Regularization

The trace norm, also called the nuclear norm, is the sum of a matrix's singular values. As a convex relaxation of matrix rank, it is commonly used in regularization to promote low-rank solutions.

Timeline (last 90 days)

Oct 19

The study introduces trace-norm regularization and proves that among all matrices producing the same observed output, the optimizer tends to choose low-rank matrices.

Unverified50%

All Facts (1)

Unverified

The study introduces trace-norm regularization and proves that among all matrices producing the same observed output, the optimizer tends to choose low-rank matrices.

50%

Source Articles

Finite-Rank Bottleneck: A New Theoretical Perspective on LoRA Low-Rank AdaptationOct 19