We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.CV

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Computer Vision and Pattern Recognition

Title: Fast Differentiable Matrix Square Root

Abstract: Computing the matrix square root or its inverse in a differentiable manner is important in a variety of computer vision tasks. Previous methods either adopt the Singular Value Decomposition (SVD) to explicitly factorize the matrix or use the Newton-Schulz iteration (NS iteration) to derive the approximate solution. However, both methods are not computationally efficient enough in either the forward pass or in the backward pass. In this paper, we propose two more efficient variants to compute the differentiable matrix square root. For the forward propagation, one method is to use Matrix Taylor Polynomial (MTP), and the other method is to use Matrix Pad\'e Approximants (MPA). The backward gradient is computed by iteratively solving the continuous-time Lyapunov equation using the matrix sign function. Both methods yield considerable speed-up compared with the SVD or the Newton-Schulz iteration. Experimental results on the de-correlated batch normalization and second-order vision transformer demonstrate that our methods can also achieve competitive and even slightly better performances. The code is available at \href{https://github.com/KingJamesSong/FastDifferentiableMatSqrt}{this https URL}.
Comments: Accpeted by ICLR 2022
Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG); Mathematical Software (cs.MS); Numerical Analysis (math.NA)
Cite as: arXiv:2201.08663 [cs.CV]
  (or arXiv:2201.08663v1 [cs.CV] for this version)

Submission history

From: Yue Song [view email]
[v1] Fri, 21 Jan 2022 12:18:06 GMT (10852kb,D)

Link back to: arXiv, form interface, contact.