AI Paper Atlas

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

Asymmetric Loss for Multi-Label Classification

Tal Ridnik, Emanuel Ben-Baruch, Nadav Zamir, Asaf Noy, Itamar Friedman, Matan Protter, Lihi Zelnik-Manor

ICCV 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Partial and multi-label learning · 867 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
replace cross-entropy with focal lossTResNet-L · MS-COCOmAP: 84.0 -> 85.1Improved
add asymmetric focusing (lower gamma+, gamma+=0) (ablation)focal loss · MS-COCOmAP: 85.1 -> 86.3Improved
add asymmetric probability margin (ablation)focal loss · MS-COCOmAP: 85.1 -> 86.3Improved
add probability margin to asymmetric focusing (ablation)ASL · MS-COCOmAP: 86.3 -> 86.6Improved
use ASL instead of static linear weighting (ablation)focal loss · MS-COCOmAP: 85.3 -> 86.6Improved
add linear positive weighting (ablation)focal loss · MS-COCOmAP: 85.1 -> 85.3No clear effect
adaptive asymmetry with unbiased target gap 0 (ablation)ASL · MS-COCOmAP: 85.1 -> 85.8Improved
raise target probability gap from 0 to 0.2 (ablation)adaptive ASL · MS-COCOmAP: 85.8 -> 86.4Improved
use adaptive instead of fixed gamma- (ablation)ASL · MS-COCOmAP: 86.6 -> 86.4Hurt
use ImageNet-21K instead of 1K pretrainingASL TResNet-L · MS-COCOmAP: 86.6 -> 88.4Improved
increase input resolution from 448 to 640ASL TResNet-L · MS-COCOmAP: 88.4 -> 89.8Improved
use TResNet-L instead of ResNet101 backboneASL · MS-COCOmAP: 85.0 -> 86.6Improved

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