AI Paper Atlas

Research map / Continual learning / Class-incremental

Decoupling the Class Label and the Target Concept in Machine Unlearning

Jianing Zhu, Bo Han, Jiangchao Yao, Jianliang Xu, Gang Niu, Masashi Sugiyama

ICLR 2026 · Continual learning › Class-incremental · approach: Class-balanced knowledge retention · 11 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
use TARF instead of gradient ascent (GA)ResNet-18 · CIFAR-10 target mismatch forgettingaverage gap: 20.80 -> 1.23Improved
use TARF instead of fine-tuning (FT)ResNet-18 · CIFAR-100 target mismatch forgettingaverage gap: 28.54 -> 0.21Improved
use TARF instead of fine-tuning (FT)ResNet-18 · CIFAR-10 model mismatch forgettingaverage gap: 5.33 -> 2.90Improved
use TARF instead of FTResNet-18 · CIFAR-100 data mismatch forgettingaverage gap: 37.15 -> 1.17Improved
use TARF instead of FTResNet-18 · CIFAR-10 all matched forgettingaverage gap: 1.07 -> 1.01No clear effect
use TARF instead of SCRUBResNet-18 · CIFAR-100 all matched forgettingaverage gap: 0.71 -> 1.11Hurt
use TARF instead of GAResNet-18 · CIFAR-100 data mismatch forgettingaverage gap: 2.43 -> 1.17Improved
use TARF instead of GAResNet-18 · run-time efficiency CIFAR-10 all matchedRTE: 0.25 -> 4.21Hurt
use annealed (learning-rate-reduced) gradient ascent instead of constant k (ablation)TARF · model mismatch forgettingcloseness to Retrained on forgetting dataImproved
use gradient cleaning (0) instead of gradient ascent on selected forgetting data (ablation)TARF · target mismatch forgettingretaining accuracyImproved
increase initial ascent strength k beyond about 0.05 (ablation)TARF · CIFAR-100 all matched forgettinggapHurt

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