AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

DiffAttack: Evasion Attacks Against Diffusion-Based Adversarial Purification

Mintong Kang, Dawn Song, Bo Li

NeurIPS 2023 · Adversarial robustness and security › Certified and empirical robustness · approach: Augmentation and diffusion purification · 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
Use DiffAttack instead of AdjAttackscore-based purification, WideResNet-28-10 · CIFAR-10 linf 8/255robust accuracy (%): 70.64 -> 46.88Improved
Use DiffAttack instead of AdjAttackscore-based purification, WideResNet-70-16 · CIFAR-10 linf 8/255robust accuracy (%): 71.29 -> 45.31Improved
Use DiffAttack instead of AdjAttackscore-based purification, WideResNet-28-10 · CIFAR-10 linf 4/255robust accuracy (%): 82.81 -> 71.88Improved
Use DiffAttack instead of AdjAttackscore-based purification, WideResNet-28-10 · CIFAR-10 l2 0.5robust accuracy (%): 78.58 -> 64.06Improved
Use DiffAttack instead of AdjAttackscore-based purification, ResNet-50 · ImageNet linf 4/255robust accuracy (%): 40.93 -> 28.13Improved
Use DiffAttack instead of Diff-BPDADDPM-based purification, WideResNet-28-10 · CIFAR-10 linf 8/255robust accuracy (%): 75.00 -> 54.69Improved
Use DiffAttack instead of Diff-BPDADDPM-based purification, WideResNet-70-16 · CIFAR-10 l2 0.5robust accuracy (%): 81.25 -> 71.88Improved
Use DiffAttack instead of SPSAscore-based purification · CIFAR-10 linf 8/255robust accuracy (%): 83.37 -> 46.88Improved
Use DiffAttack instead of Square AttackDDPM-based purification · CIFAR-10 linf 8/255robust accuracy (%): 81.68 -> 54.69Improved
Use DiffAttack instead of Joint Attack (Full)DDPM-based purification · CIFAR-10 linf 8/255robust accuracy (%): 76.26 -> 54.69Improved
Use segment-wise forwarding-backwarding instead of full computational graph (ablation)gradient backpropagation through diffusion · CIFAR-10 WideResNet-28-10, batch 16memory costImproved
Increase diffusion length T of the purifier (ablation)diffusion-based purification · CIFAR-10 WideResNet-28-10clean accuracyHurt

Later papers that compared with this method

11 later accepted papers report DiffAttack as a baseline in their results tables; 8 of them report a better number on most of the results they share with it.

3 more not listed.

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