AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Few-Shot Backdoor Attacks on Visual Object Tracking

Yiming Li, Haoxiang Zhong, Xingjun Ma, Yong Jiang, Shu-Tao Xia

ICLR 2022 · Adversarial robustness and security › Backdoor and poisoning · approach: Backdoors in large multimodal models · 21 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 feature-space loss maximization alongside tracking-loss minimization instead of BOBA label flippingSiamFC++ · OTB100, one-shotAUC-A: 57.79 -> 18.78Improved
Use feature-space loss maximization alongside tracking-loss minimization instead of BOBA label flippingSiamFC++ · GOT10K validation, one-shotAUC-A: 66.70 -> 35.82Improved
Use feature-space loss maximization alongside tracking-loss minimization instead of BOBA label flippingSiamFC++ · LaSOT, one-shotAUC-A: 43.06 -> 16.39Improved
Increase the poisoning rate from 0% to 10% (ablation)FSBA on SiamFC · OTB100, one-shotPr-A: 72.43 -> 11.06Improved
Increase the poisoning rate from 0% to 10% (ablation)FSBA on SiamFC · OTB100, few-shotPr-A: 74.03 -> 7.92Improved
Increase the poisoning rate from 0% to 10% (ablation)FSBA on SiamFC · OTB100Pr-B: 79.23 -> 75.98Hurt
Train SiamFC++ with additional GOT10K data (ablation)FSBA · OTB100, one-shotAUC-A: 18.78 -> 23.20Hurt
Train SiamFC++ with additional LaSOT data (ablation)FSBA · OTB100, one-shotAUC-A: 23.20 -> 23.19No clear effect
Apply frame-wise color jittering as a defense (ablation)FSBA SiamFC++ · OTB100, one-shotPr-A: Pr-A is still below 40% in all casesNo clear effect
Fine-tune using benign samples (ablation)FSBA SiamFC++ · OTB100, one-shotPr-A: performance gap ... is still larger than 40% in all casesNo clear effect
Add Gaussian noise to test frames (ablation)FSBA SiamFC++ · OTB100, one-shotPr-A: <30%No clear effect
Prune model channels at rates from 0% to 30% (ablation)FSBA SiamFC++ · OTB100, one-shotPr-A: 25.52; 26.70; 27.40; 27.42; 29.04; 25.29; 21.69No clear effect

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