AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

AdvFlow: Inconspicuous Black-box Adversarial Attacks using Normalizing Flows

Hadi Mohaghegh Dolatabadi, Sarah Erfani, Christopher Leckie

NeurIPS 2020 · Adversarial robustness and security › Black-box and transfer attacks · approach: Robustness benchmarks and defenses · 19 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 a clean-data-pretrained normalizing flow instead of N-Attack's search distributionblack-box attack · CIFAR-10 defended classifiersattack success rate: 38.97 -> 41.21; 36.90 -> 40.22; 38.04 -> 40.67Improved
Use a clean-data-pretrained normalizing flow instead of N-Attack's search distributionblack-box attack · SVHN defended classifiersattack success rate: 50.28 -> 50.76; 35.42 -> 41.49; 41.49 -> 44.22Improved
Use AdvFlow rather than N-Attackblack-box attack · CIFAR-10 vanilla classifierattack success rate: 100 -> 99.42Hurt
Use AdvFlow rather than N-Attackblack-box attack · SVHN vanilla classifierattack success rate: 98.76 -> 90.31Hurt
Pretrain the flow on clean data rather than use random flow weights (ablation)AdvFlow · CIFAR-10, LID detectordetection AUROC: 84.39 -> 57.59Improved
Pretrain the flow on clean data rather than use random flow weights (ablation)AdvFlow · CIFAR-10, Mahalanobis detectordetection AUROC: 99.50 -> 66.85Improved
Pretrain the flow on clean data rather than use random flow weights (ablation)AdvFlow · CIFAR-10, Res-Flow detectordetection AUROC: 99.40 -> 67.03Improved
Pretrain the flow on clean data rather than use random flow weights (ablation)AdvFlow · SVHN, LID detectordetection AUROC: 58.92 -> 61.11Hurt
Pretrain the flow on clean data rather than use random flow weights (ablation)AdvFlow · SVHN, Mahalanobis detectordetection AUROC: 74.67 -> 64.72Improved
Pretrain the flow on clean data rather than use random flow weights (ablation)AdvFlow · SVHN, Res-Flow detectordetection AUROC: 74.86 -> 64.68Improved
Use AdvFlow rather than N-Attackblack-box attack · CIFAR-10 defended classifiersaverage query count: 874.91 -> 421.63; 973.05 -> 436.8; 941.67 -> 424.95Mixed
Use AdvFlow rather than N-Attackblack-box attack · SVHN defended classifiersaverage query count: 903.12 -> 692.73; 1243.38 -> 526.37; 756.48 -> 480.02Mixed

Later papers that compared with this method

1 later accepted paper reports AdvFlow as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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