AI Paper Atlas

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

Learning Hierarchy Aware Features for Reducing Mistake Severity

Ashima Garg, Depanshu Sani, Saket Anand

ECCV 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Partial and multi-label learning · 22 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
add geometric consistency loss L_gc to soft-consistency+margin (ablation)HAF · CIFAR-100mistake severity: 2.24 -> 2.22Improved
add margin loss L_m to soft consistency + geometric consistency (ablation)HAF · CIFAR-100mistake severity: 2.32 -> 2.22Improved
add margin loss L_m to soft consistency + geometric consistency (ablation)HAF · CIFAR-100hierarchical distance@5: 1.66 -> 1.40Improved
add soft hierarchical consistency L_shc to cross-entropy (ablation)HAF · CIFAR-100hierarchical distance@5: 2.24 -> 1.61Improved
add soft hierarchical consistency L_shc to cross-entropy (ablation)HAF · CIFAR-100top-1 error (%): 22.11 -> 22.70Hurt
add L_gc to soft consistency + margin loss (ablation)HAF · CIFAR-100top-1 error (%): 22.12 -> 22.25Hurt
add full HAF losses to cross-entropy (ablation)HAF · iNaturalist-19mistake severity: 2.39 -> 2.27Improved
add full HAF losses to cross-entropy (ablation)HAF · iNaturalist-19top-1 error (%): 36.48 -> 36.47No clear effect
add L_gc to soft consistency alone (ablation)HAF · iNaturalist-19top-1 error (%): 36.23 -> 36.60Hurt
train with HAF instead of cross-entropyCross-Entropy · CIFAR-100mistake severity: 2.35 -> 2.24Improved
train with HAF instead of cross-entropyCross-Entropy · tieredImageNet-Hmistake severity: 7.05 -> 7.03No clear effect
apply CRM test-time post-processing to HAFHAF · CIFAR-100mistake severity: 2.24 -> 2.23No clear effect

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