AI Paper Atlas

Research map / Learning theory and algorithms / Matrix and tensor methods

Fast Differentiable Matrix Square Root and Inverse Square Root

Yue Song, Nicu Sebe, Wei Wang

TPAMI 2022 · Learning theory and algorithms › Matrix and tensor methods · approach: Low-rank approximation theory · 40 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
increase MPA forward degree from [3,3] to [5,5] (ablation)MPA-Lya · CIFAR10 ZCA whitening (ResNet-18)validation error (%): 4.64 -> 4.39Improved
increase MPA forward degree from [5,5] to [6,6] (ablation)MPA-Lya · CIFAR10 ZCA whitening (ResNet-18)validation error (%): 4.39 -> 4.45No clear effect
replace exact Bartels-Stewart backward with iterative Lyapunov solver (8 iterations) (ablation)MPA-Lya · CIFAR10 ZCA whitening (ResNet-18)validation error (%): 4.57 -> 4.39Improved
increase Lyapunov backward iterations from 5 to 8 (ablation)MPA-Lya · CIFAR10 ZCA whitening (ResNet-18)validation error (%): 4.48 -> 4.39Improved
increase Lyapunov backward iterations from 8 to 9 (ablation)MPA-Lya · CIFAR10 ZCA whitening (ResNet-18)validation error (%): 4.39 -> 4.40No clear effect
use Pade approximant (MPA) instead of Taylor polynomial (MTP) in the forward passLyapunov backward · CIFAR100 ZCA whitening, ResNet-50validation error (%): 20.55 -> 19.55Improved
use MPA-Lya instead of NS iterationZCA whitening · CIFAR100 ZCA whitening, ResNet-50validation error (%): 19.39 -> 19.55Hurt
use MPA-Lya instead of NS iteration for covariance square rootGCP ResNet-50 · ImageNettime FP+BP (ms): 164.43 -> 110.61Improved
use MPA-Lya instead of NS iterationtemporal-attentive GCP · HMDB51top-1 accuracy (%): 72.75 -> 74.05Improved
use MPA-Lya instead of NS iterationSo-ViT-7 · ImageNettop-1 accuracy (%): 76.50 -> 76.84Improved
use MPA-Lya instead of SVD-Padéfine-grained GCP · Carsaccuracy (%): 92.8 -> 92.5Hurt
use MPA-Lya instead of NS iterationWCT style transfer · ArtworksLPIPS: 0.5578 -> 0.5615Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy