AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

Uncovering Adversarial Risks of Test-Time Adaptation

Tong Wu, Feiran Jia, Xiangyu Qi, Jiachen T. Wang, Vikash Sehwag, Saeed Mahloujifar, Prateek Mittal

ICML 2023 · Adversarial robustness and security › Attacks on text and graphs · approach: Membership inference and inversion · 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
Inject malicious test-batch samples using DIATeBN, TENT, and pseudo-label TTA · CIFAR-10-C, CIFAR-100-C, and ImageNet-Ctargeted attack success rate: ImageNet-C: 10 malicious samples (5%) yields 99.47 (TeBN), 98.67 (TENT), and 96.53 (Hard PL)Improved
Increase the number of malicious samples in the batchDIA on TTA · CIFAR-10-C, CIFAR-100-C, and ImageNet-Ctargeted attack success rate: CIFAR-10-C TeBN: 25.87 -> 55.47 -> 92.80; CIFAR-100-C TeBN: 46.80 -> 93.73 -> 100.00Improved
Inject malicious samples for an indiscriminate attackTeBN on benign corrupted samples · CIFAR-10-Ccorruption error rate: 10.73 -> 28.02Hurt
Inject malicious samples for an indiscriminate attackTeBN on benign corrupted samples · CIFAR-100-Ccorruption error rate: 34.52 -> 58.41Hurt
Inject 20 malicious samples for an indiscriminate attackTeBN on benign corrupted samples · ImageNet-Ccorruption error rate: 47.79 -> 79.03Hurt
Use adversarially optimized snow or fog corruptions as injected samplesTeBN · ImageNet-Ctargeted attack success rate: Snow: 64.00; Fog: 84.00Improved
Use adversarially trained source modelsTeBN on ImageNet-C · ImageNet-Ctargeted attack success rate: With 8 malicious samples, ASR degrades by ~80%; with 32 malicious samples, DIA still exceeds 70% successHurt
Blend training-time BN statistics with test-time BN statistics (τ = 0.6)TTA BN estimation on ImageNet-C · ImageNet-Ctargeted attack success rate: ASR to ~20% for 20 malicious samples with robust modelsHurt
Use training-time BN for the last 8 layers, with τ = 0.6 elsewhereTTA BN estimation on ImageNet-C · ImageNet-Ctargeted attack success rate: ASR drops ~40% for the standard model with N_tr = 8Hurt
Combine the mitigation approaches: adversarially trained model and robust BN estimationTTA on ImageNet-C · ImageNet-Ctargeted attack success rate: mitigate ASR by ~50% for the standard model and ~40% for the robust modelHurt
Increase the weight ω to 0.1 in the stealthy targeted objective (ablation)DIA on TeBN · CIFAR-10-Ctargeted attack success rate: ASR remains more than 75%; corruption accuracy degradation drops to ~2%Mixed
Use stronger data augmentation (AugMix or DeepAugment) (ablation)DIA on ImageNet-C · ImageNet-Ctargeted attack success rate: TeBN: 99.47 (Standard), 98.40 (AugMix), 96.00 (DeepAugment)Hurt

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