AI Paper Atlas

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

HSFM: Hard-Set-Guided Feature-Space Meta-Learning for Robust Classification under Spurious Correlations

Aryan Yazdan Parast, Khawar Islam, Soyoun Won, Basim Azam, NAVEED AKHTAR

ECCV 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Data augmentation and generalization bias · 1 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
apply HSFM on top of ERM (ResNet-50)ERM · Waterbirdsworst-group accuracy: 74.6 -> 93.1Improved
apply HSFM on top of ERM (ResNet-50)ERM · CelebAworst-group accuracy: 30.6 -> 89.2Improved
apply HSFM on top of ERM (ResNet-50)ERM · Dominoesworst-group accuracy: 78.6 -> 90.4Improved
apply HSFM on top of ERM (ResNet-50)ERM · MetaShiftworst-group accuracy: 64.1 -> 77.2Improved
apply HSFM on top of ERM (ViT-B/16)ERM · Waterbirdsworst-group accuracy: 53.9 -> 81.5Improved
apply HSFM on top of ERM (ViT-B/16)ERM · MetaShiftworst-group accuracy: 69.7 -> 69.9No clear effect
apply HSFM on top of ERM (ConvNeXt)ERM · CelebAworst-group accuracy: 47.2 -> 81.1Improved
apply HSFM on top of ERM (ResNet-50)ERM · CelebAaverage accuracy: 95.83 -> 90.6Hurt
apply HSFM on CLIP linear classifierERM · CelebA*worst-group accuracy: 50.0 -> 81.0Improved
apply HSFM on pretrained ResNet-50 featuresERM · Stanford Carstop-1 accuracy: 83.98 -> 85.10Improved
increase number of inner-loop updates T from 1 or 5 to 10 or more (ablation)HSFM · Waterbirdsworst-group accuracyImproved
use intermediate support set size (250) rather than larger (ablation)HSFM · Waterbirdsworst-group accuracyImproved

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