AI Paper Atlas

Research map / Architectures and efficiency / Quantization

Robust Machine Unlearning for Quantized Neural Networks via Adaptive Gradient Reweighting with Similar Labels

Yujia Tong, Yuze Wang, Jingling Yuan, Chuang Hu

ICCV 2025 · Architectures and efficiency › Quantization · approach: Post-training weight quantization · 7 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 Similar Labels (replace random labels for forgotten data) (ablation)Q-MUL · CIFAR-100 10% (ResNet18)Average Gap: 4.10 -> 3.11Improved
add Similar Labels (ablation)Q-MUL · CIFAR-100 10% (ResNet18)RA gap to Retrain: 3.52 -> 2.09Improved
add Adaptive Gradient Reweighting (ablation)Q-MUL · CIFAR-100 10% (ResNet18)Average Gap: 4.57 -> 3.11Improved
add Adaptive Gradient Reweighting (ablation)Q-MUL · CIFAR-100 10% (ResNet18)FA gap to Retrain: 4.92 -> 0.95Improved
use Q-MUL instead of SalUnResNet18 4-bit · CIFAR-100 10% random forgettingAverage Gap: 6.05 -> 3.11Improved
use Q-MUL instead of SalUnResNet18 4-bit · CIFAR-10 10% random forgettingAverage Gap: 1.31 -> 1.30No clear effect
use Q-MUL instead of SalUnResNet18 4-bit · CIFAR-100 50% random forgettingAverage Gap: 7.33 -> 7.06Improved
use Q-MUL instead of SalUnResNet18 PACT QAT · CIFAR-100 10%Average Gap: 9.26 -> 6.00Improved
use Q-MUL instead of RLResNet18 DSQ QAT · CIFAR-100 10%Average Gap: 9.88 -> 8.26Improved
use Q-MUL instead of l1-sparseResNet18 4-bit · Tiny-ImageNet 10%Average Gap: 9.02 -> 8.17Improved
use Q-MUL instead of RL/l1-sparse/SalUn in class-wise forgettingResNet18 4-bit · CIFAR-10 class-wiseAverage Gap: 0.13, 0.08, 0.10 vs 0.11No clear effect
use Q-MUL instead of SalUnMobileNetV2 2-bit · CIFAR-100 10%Average Gap: 9.29 -> 6.52Improved

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