AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Rotation Regularization without Rotation

Takumi Kobayashi

ECCV 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 0 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
use statistical rotation instead of geometric rotation matrices (ablation)rotation regularization · ImageNet-LTerror rate (%): 56.68 -> 56.40Improved
use correlated random unit vector instead of i.i.d. Gaussian noise (ablation)StatRot · ImageNet-LTerror rate (%): 56.40 -> 56.41No clear effect
add the random rotation noise term on top of plain logit scaling (ablation)StatRot · ImageNet-LTerror rate (%): 58.25 -> 56.40Improved
use angular cone instead of Euclidean margin ball (ablation)StatRot · ImageNet-LTerror rate (%): 56.94 -> 56.40Improved
make noise asymmetric (target logit only, -|epsilon|) (ablation)StatRot · ImageNet-LTerror rate (%): 56.40 -> 56.45No clear effect
combine StatRot with DropOut (ablation)StatRot · ImageNet-LTerror rate (%): 56.40 -> 56.12Improved
use theoretical D = min(d,C) = 512 instead of smaller D = 128 (ablation)StatRot · ImageNet-LTerror rate (%)Improved
set rotation angle to pi/4 rather than larger angles (ablation)StatRot · ImageNet-LTerror rate (%)Improved
add StatRot to softmax trainingResNet10 · ImageNet-LTerror rate (%): 58.91 -> 56.40Improved
add StatRot to softmax trainingResNet50 · iNat2018error rate (%): 32.82 -> 30.39Improved
add StatRot to softmax trainingResNet10 · Places-LTerror rate (%): 72.99 -> 71.85Improved
add StatRot to softmax trainingResNet152 pretrained · Places-LTerror rate (%): 61.14 -> 59.93Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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