AI Paper Atlas

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

Tackling Fake Forgetting through Uncertainty Quantification

Yingdan Shi, Sijia Liu, Kaize Ding, Ren Wang

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Confidence calibration objectives · 0 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
add the CPU conformal loss (lambda 0 to 0.5) (ablation)FT · CIFAR-10 ResNet-18, 10% random forgettingUA gap to RT: 4.8% -> 0.7%Improved
add the CPU conformal loss (lambda 0 to 0.5) (ablation)FT · Tiny ImageNet ViT, 10% random forgettingUA gap to RT: 7.8% -> 0.9%Improved
add the CPU conformal loss (lambda 0 to 0.5) (ablation)FT · CIFAR-10 ResNet-18, 10% random forgettingaverage gap across UA, RA, TA: 2.2 -> 1.47Improved
add the CPU conformal loss (lambda 0 to 0.5) (ablation)FT · Tiny ImageNet ViT, 10% random forgettingaverage gap across UA, RA, TA: 3.53 -> 1.63Improved
add the CPU conformal loss (lambda 0 to 1.0) (ablation)FT · CIFAR-10 ResNet-18, 10% random forgettingCR on forget data: 0.986 -> 0.835Improved
add the CPU conformal loss (lambda 0 to 1.0) (ablation)RL · CIFAR-10 ResNet-18, 10% random forgettingCR on forget data: 0.788 -> 0.629Improved
add the CPU conformal loss (lambda 0 to 1.0) (ablation)RT · CIFAR-10 ResNet-18, 10% random forgettingUA: 8.6% -> 17.7%Improved
add the CPU conformal loss (lambda 0 to 1.0) (ablation)RT · Tiny ImageNet ViT, 10% random forgettingUA: 14.7% -> 35.7%Improved
add the CPU conformal loss (lambda 0 to 1.0) (ablation)RL · CIFAR-10 ResNet-18, 10% random forgettingretain accuracy RA: 97.4% -> 95.3%Hurt
add the CPU conformal loss (lambda 0 to 0.5), averaged over methods (ablation)all methods · CIFAR-10 ResNet-18 and ViTtest accuracy TA: decreases only slightly by 1.0% and 0.57%Hurt
add the CPU conformal loss (lambda 0 to 0.5), averaged over methods (ablation)all methods · CIFAR-10 ResNet-18 and ViTCR on test data: 0.03 drop on averageNo clear effect
add the CPU conformal loss (lambda 0 to 0.5), averaged over methods (ablation)all methods · CIFAR-10 ResNet-18 and ViTCR on forget data: 0.08 average reductionImproved

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