New answers tagged

1 vote

Cheap way to keep parameter matrices orthogonal during optimization?

There is something related in: Bujanović, Zvonimir; Kressner, Daniel; Schröder, Christian, Iterative refinement of Schur decompositions, Numer. Algorithms 92, No. 1, 247-267 (2023). ZBL1507.65076. ...
Thijs Steel's user avatar
  • 1,568
0 votes

Cheap way to keep parameter matrices orthogonal during optimization?

It appears a method based on Householder reflectors (background) is used in PyTorch for their orthogonal parameterization link project original (X) to orthogonal space (Q) at each step: Using ...
Yaroslav Bulatov's user avatar
4 votes

Weird runtime behavior of `scipy.linalg.solve_triangular` and `trtrs`

Thanks to Federico Poloni's comment, I have revised the script as below. It looks like scipy.linalg.solve_triangular has time complexity $O(p^2)$, because the ...
nalzok's user avatar
  • 181
8 votes

Saddle point system

$$\newcommand{\img}{\operatorname{img}}$$ You mention this: Are there other efficient ways of solving $Ax=c$ when $\dim\ker A=1$ and an arbitrary one dimensional orthogonality constraint is ...
lightxbulb's user avatar
  • 1,994

Top 50 recent answers are included