AI Paper Atlas

Research map / Optimization / Optimizers

Learning to Unlearn while Retaining: Combating Gradient Conflicts in Machine Unlearning

Gaurav Patel, Qiang Qiu

ICCV 2025 · Optimization › Optimizers · approach: Surrogate losses and robust solvers · 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 the inner retain-gradient step (LUR objective) instead of plainly summing retain and forget losses (ablation)LUR · CIFAR-10 random forgetting 10%Avg. Gap to Retrain: 3.15 -> 0.89Improved
add the inner retain-gradient step (LUR objective) instead of plainly summing retain and forget losses (ablation)LUR · CIFAR-10 random forgetting 50%Avg. Gap to Retrain: 3.51 -> 2.62Improved
add the inner retain-gradient step (LUR objective) instead of plainly summing retain and forget losses (ablation)LUR · CIFAR-100 random forgetting 10%Avg. Gap to Retrain: 4.28 -> 3.96Improved
add the inner retain-gradient step (LUR objective) instead of plainly summing retain and forget losses (ablation)LUR · CIFAR-100 random forgetting 50%Avg. Gap to Retrain: 15.64 -> 8.07Improved
use LUR instead of SHs, explicit gradient projection, for identity forgettingResNet-34 · Celeb-HQ-FIR identity forgetting 10%Avg. Gap to Retrain: 2.82 -> 0.10Improved
use LUR instead of SalUn for 50% random forgetting on CIFAR-10ResNet-18 · CIFAR-10 random forgetting 50%Avg. Gap to Retrain: 2.18 -> 2.62Hurt
use LUR instead of SalUn for CIFAR-100 forgettingResNet-18 · CIFAR-100 random forgetting 10%Avg. Gap to Retrain: 10.45 -> 3.96Improved
use LUR instead of SA for DDPM class unlearningDDPM · CIFAR-10 class-wise forgetting (DDPM)UA: 85.80 -> 100.00Improved
use LUR instead of SA for DDPM class unlearningDDPM · CIFAR-10 class-wise forgetting (DDPM)FID: 9.08 -> 9.76Hurt
use LUR instead of SalUn for Stable Diffusion class unlearningStable Diffusion · Imagenette class-wise forgetting (average)FID: 1.22 -> 0.98Improved
vary the inner learning step alpha; too large or too small (ablation)LUR · Classification unlearningAvg. Gap to Retrain: optimal alpha = 0.01Mixed

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