AI Paper Atlas

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

SLIM: Spuriousness Mitigation with Minimal Human Annotations

Xiwei Xuan, Ziquan Deng, Hsuan-Tien Lin, Kwan-Liu Ma

ECCV 2024 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory · 7 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
curate a feature-balanced subset with SLIM instead of plain ERMResNet50 · Waterbirdsworst-group accuracy (%): 62.6 -> 89.1 (SLIM_Tr), 91.0 (SLIM_Val)Improved
use SLIM_Tr with attention labels instead of CNC with extra trainingResNet50 · ImageNet-9worst-group accuracy (%): 58.9 -> 68.1Improved
use SLIM_Tr instead of GDRO with full spuriousness labelsResNet50 · CelebAworst-group accuracy (%): 88.9 -> 89.4No clear effect
use attention-correctness labels instead of spuriousness labelshuman annotation · NICO labeling studyannotation consistency (%): 70.1 -> 95.1Improved
increase number of attention annotations (ablation)SLIM_Tr · Waterbirdsworst-group accuracy (%): 78.2 -> 89.1 (60 -> 120 annotations)Improved
increase annotations beyond 120 (ablation)SLIM_Tr · Waterbirdsworst-group accuracy (%): 89.1 -> 89.4 (120 -> 150)No clear effect
enlarge constructed training subset beyond the chosen size (ablation)SLIM_Tr · Waterbirdsworst-group accuracy (%): 89.1 -> 83.5 (10% -> 15%)Hurt
use attention space (attention-weighted features) instead of original feature space (ablation)SLIM label propagation · Waterbirdsneighbor attention consistencyImproved
apply SLIM_Tr on ISIC with attention labelsResNet50 · ISICworst-group accuracy (%): 56.7 -> 79.1Improved
apply SLIM to improve attention accuracyResNet50 · Waterbirdsworst-group AIoU: .25 -> .53Improved

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