AI Paper Atlas

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

Anchor Loss: Modulating Loss Scale Based on Prediction Difficulty

Serim Ryou, Seong-Gyun Jeong, Pietro Perona

ICCV 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory · 45 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 anchor lossResNet-110 · CIFAR-100top-1 accuracy: 72.98 -> 74.25Improved
replace binary cross entropy with anchor lossResNet-110 · CIFAR-100top-1 accuracy: 73.88 -> 74.25Improved
replace cross entropy with anchor lossResNet-50 · ImageNet valtop-1 accuracy: 76.39 -> 76.82Improved
use OHEM hard-example sampling instead of CEResNet-110 · CIFAR-100top-1 accuracy: 72.98 -> 73.03No clear effect
add CE warmup for first 5 epochs to anchor lossResNet-110 anchor loss · CIFAR-100top-1 accuracy: 74.25 -> 74.38Improved
use network-derived dynamic anchor instead of static q*=0.5 (ablation)Anchor loss · CIFAR-100top-1 accuracy: 73.77 -> 74.25Improved
raise gamma from 0.5 to 2.0 (ablation)Anchor loss · CIFAR-100top-1 accuracy: 74.25 -> 71.86Hurt
modulate background loss (ii) instead of target loss (i) (ablation)Anchor loss · CIFAR-100top-1 accuracy: 74.06 -> 74.25Improved
modulate both target and background (iii) instead of background only (ablation)Anchor loss · CIFAR-100top-1 accuracy: 74.25 -> 73.90Hurt
replace MSE with anchor lossHourglass · MPII testPCK total: 90.9 -> 91.9Improved
replace MSE with anchor loss, single-scaleHourglass · MPII valPCK mean: 89.32 -> 89.93Improved
set gamma=2 instead of gamma=5 in anchor loss (ablation)2-stack Hourglass · MPII valPCK mean: 88.84 -> 89.11Improved

Later papers that compared with this method

1 later accepted paper reports Anchor Loss as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

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