AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Sharpness-Aware Machine Unlearning

Haoran Tang, Rajiv Khanna

ICLR 2026 · Optimization › Implicit bias and generalization · approach: Sharpness-aware minimization theory · 5 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 ASAM/SAM as the unlearning optimizer instead of SGD (ablation)existing unlearning methods (NegGrad, RL, SalUn, SCRUB, L1-Sparse) · CIFAR-100 and ImageNet-1KToW: SCRUB 49.084 -> 73.766 (CIFAR-100, SGD pretrain)Improved
pretrain with SAM/ASAM instead of SGD (ablation)NegGrad unlearning · CIFAR-100ToW: 81.796 -> 84.789 (ASAM 0.1), 81.796 -> 72.78 (SAM 0.1)Mixed
use Sharp MinMax (SAM on retain weights, sharpness maximization on forget weights)unlearning with SAM 0.1 · CIFAR-100ToW: 83.514 -> 91.344Improved
use Sharp MinMax instead of NegGrad with ASAM 1.0ImageNet-1K SGD-pretrained · ImageNet-1KToW: 84.843 -> 87.899Improved
apply SAM/ASAM unlearning (vs SGD) to reduce MIA correctness on forget set (ablation)L1-Sparse, SCRUB, RL, SalUn, NegGrad · CIFAR-100MIA correctness: SCRUB 35.544 -> 20.478Improved
use SAM instead of SGD for unlearning (feature entanglement) (ablation)NegGrad · CIFAR-100E_Wp entanglement: 53.45 -> 52.39Improved
use SAM instead of SGD for unlearning (variance entanglement) (ablation)NegGrad · CIFAR-100E_Var entanglement: 26.05 -> 26.58No clear effect
reduce NegGrad retain weight alpha (ablation)SGD vs SAM/ASAM unlearning · CIFAR-100ToWHurt
use SAM vs SGD for unlearning: forget accuracy (ablation)NegGrad · CIFAR-100 / ImageNet-1Kforget accuracyHurt
add SAM in unlearning on ViT-Small with AdamW (ablation)NegGrad and Sharp MinMax · CIFAR-100 (ViT-Small)ToWImproved

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