AI Paper Atlas

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

RankMixup: Ranking-Based Mixup Training for Network Calibration

Jongyoun Noh, Hyekang Park, Junghyup Lee, Bumsub Ham

ICCV 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 14 citations (OpenAlex)

Read the paper

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
increase MRL margin m beyond 2.0 on CIFAR10 (ablation)RankMixup (MRL) · CIFAR10ECE (%): 1.01 -> 13.13 (m=2.0 -> 5.0)Hurt
increase number of augmented samples Q in M-NDCG (ablation)RankMixup (M-NDCG) · Tiny-ImageNetECE (%): 3.25 -> 1.30 (Q=2 -> Q=6)Improved
increase number of augmented samples Q in M-NDCG (ablation)RankMixup (M-NDCG) · CIFAR10ECE (%): 3.01 -> 2.24 (Q=2 -> Q=6)Improved
use flat Beta(alpha,alpha) with large alpha (2 on CIFAR10, 1 on Tiny-ImageNet) (ablation)RankMixup (M-NDCG) · CIFAR10, Tiny-ImageNetECE/AECEImproved
replace vanilla mixup with Manifold-Mixup in RankMixup (ablation)RankMixup (M-NDCG) · Tiny-ImageNetECE (%): 1.49 -> 1.20Improved
replace vanilla mixup with Manifold-Mixup in RankMixup (ablation)RankMixup (M-NDCG) · CIFAR10ECE (%): 2.59 -> 2.92Hurt
replace label-mixture supervision with RankMixup ranking losses vs RegMixup variantsRankMixup with Manifold-Mixup/CutMix · Tiny-ImageNetECE (%): 4.76 -> 1.20 (Manifold-Mixup)Improved
apply TS post-hoc scaling on top of RankMixupRankMixup (M-NDCG) · CIFAR10 (ResNet-50)ECE (%): 2.59 -> 0.57Improved
use M-NDCG instead of MRL on long-tailed data (ablation)RankMixup · CIFAR100-LT rho=100ECE (%): 5.28 -> 18.74Hurt
use RankMixup ranking losses instead of mixup label mixturesMixup baseline · CIFAR100-LT rho=10ECE (%): 19.69 -> 9.99Improved
use M-NDCG instead of MRL (ablation)RankMixup · ImageNet (ResNet-50)ECE (%): 5.70 -> 3.93Improved
train with M-NDCG for OOD detection vs mixupRankMixup (M-NDCG) · OOD CIFAR10 ID vs SVHNAUROC (%): 79.72 -> 92.62Improved

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