AI Paper Atlas

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

Towards Calibrated Hyper-Sphere Representation via Distribution Overlap Coefficient for Long-Tailed Learning

Hualiang Wang, Siming Fu, Xiaoxuan He, Hangxiang Fang, Zuozhu Liu, Haoji Hu

ECCV 2022 · Label-efficient and robust learning › Recognition and long-tail · approach: Decoupled and calibrated long-tail training · 15 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 L_icd inter-class discrepancy loss (ablation)vMF classifier · ImageNet-LTfew-shot accuracy: 26.9 -> 31.7Improved
add L_icd inter-class discrepancy loss (ablation)vMF classifier · ImageNet-LToverall accuracy: 52.2 -> 52.9Improved
add L_cfc class-feature consistency loss (ablation)vMF classifier · ImageNet-LToverall accuracy: 52.2 -> 53.1Improved
add L_cfc class-feature consistency loss (ablation)vMF classifier · ImageNet-LTfew-shot accuracy: 26.9 -> 27.9Mixed
add both L_icd and L_cfc (ablation)vMF classifier · ImageNet-LToverall accuracy: 52.2 -> 53.5Improved
use cosine baseline replaced by vMF classifier with both losses (ablation)cosine classifier · ImageNet-LToverall accuracy: 51.8 -> 53.5Improved
raise loss weight lambda from 0.2 to 0.4 (ablation)vMF classifier + both losses · ImageNet-LToverall accuracy: 53.5 -> 52.6Hurt
lower loss weight lambda from 0.2 to 0.1 (ablation)vMF classifier + both losses · ImageNet-LToverall accuracy: 53.5 -> 53.2Hurt
add post-training calibration to linear classifier (ablation)linear classifier · ImageNet-LToverall accuracy: 43.2 -> 48.3Improved
add post-training calibration to tau-norm classifier (ablation)tau-norm classifier · ImageNet-LToverall accuracy: 48.6 -> 53.0Improved
add post-training calibration to causal classifier (ablation)causal classifier · ImageNet-LToverall accuracy: 49.0 -> 50.9Improved
add post-training calibration to vMF classifier (ablation)vMF classifier · ImageNet-LToverall accuracy: 53.5 -> 53.7Improved

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