AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Injection, Attack and Erasure: Revocable Backdoor Attacks via Machine Unlearning

Baogang Song, Dongdong Zhao, Jianwen Xiang, Qiben Xu, Zizhuo Yu

AAAI 2026 · Adversarial robustness and security › Backdoor and poisoning · approach: Trigger-based backdoor attacks · 0 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 unlearning loss to trigger optimization (ablation)revocable backdoor trigger generator · CIFAR-10 (First-Order)ASR after unlearning: 32.52 -> 25.12Improved
add fixed sample partition and PCGrad (ablation)revocable backdoor trigger generator · CIFAR-10 (First-Order)ASR after unlearning: 68.60 -> 25.12Improved
add fixed sample partition and PCGrad (ablation)revocable backdoor trigger generator · CIFAR-10ASR: 98.57 -> 98.36No clear effect
add fixed sample partition and PCGrad (ablation)gradient conflict · CIFAR-10cosine similarity of gradients: -0.65 -> -0.35Improved
use revocable-trigger optimization vs COMBATbackdoor attack · CIFAR-10 (First-Order)ASR after unlearning: 45.97 -> 25.12Improved
use revocable-trigger optimization vs BadNetsbackdoor attack · CIFAR-10ASR: 29.8 -> 98.36Improved
raise poisoning rate from 0.5% to 5% (ablation)revocable backdoor · CIFAR-10ASR: 69.73 -> 98.36Improved
raise poisoning rate from 0.5% to 5% (ablation)revocable backdoor · CIFAR-10ASR after unlearning: 5.34 -> 25.12Hurt
raise trigger strength eta from 0.04 to 0.12 (ablation)revocable backdoor · CIFAR-10ASR: 71.82 -> 99.91Improved
raise trigger strength eta from 0.04 to 0.12 (ablation)revocable backdoor · CIFAR-10ASR after unlearning: 9.71 -> 44.40Hurt
unlearning with UnrollSGD instead of First-Orderrevocable backdoor · CIFAR-10 / ImageNet-10benign accuracy after unlearning: 90.05 -> 82.93Mixed
apply fine-pruning defenserevocable backdoor · CIFAR-10 (First-Order)ASRNo clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 3 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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