AI Paper Atlas

Research map / Adversarial robustness and security / Backdoor and poisoning

Capacity Overflow: A Blind Spot for Backdoor Attacks in Vision MoE

Xiaocheng Zou, Tiancheng Zheng, Xiaolin Xu, Ruyi Ding

ECCV 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
use ascending token dispatch order instead of standard order at cf=0.3 (V-MoE) (ablation)Capacity Overflow · ImageNet-100ASR (%): 69.4 -> 87.0Improved
use ascending token dispatch order instead of standard order at cf=0.3 (Swin-MoE) (ablation)Capacity Overflow · ImageNet-100ASR (%): 42.0 -> 83.6Improved
ascending dispatch order at nominal cf=1.05 (ablation)Capacity Overflow · ImageNet-100ASR (%): 2.7 -> 4.5No clear effect
lower capacity factor from 1.05 to 0.3 (ablation)Capacity Overflow · ImageNet-100 (V-MoE, B=64, standard dispatch)ASR (%): 2.7 -> 69.4Improved
place neutralizer at deeper layer layer_7 instead of layer_5 (V-MoE) (ablation)Capacity Overflow · ImageNet-100ASR activation mode (%): 86.8 -> 93.8Improved
place neutralizer at deepest layer layer_9 instead of layer_7 (V-MoE) (ablation)Capacity Overflow · ImageNet-100dormant ASR (%): 2.8 -> 4.2Hurt
place neutralizer at s2_b5 instead of s2_b1 (Swin-MoE) (ablation)Capacity Overflow · ImageNet-100ASR activation mode (%): 84.1 -> 96.8Improved
attack with Capacity Overflow vs unattacked model on clean accuracyV-MoE · ImageNet-100clean accuracy (%): 76.0% -> 74.7%Hurt
Capacity Overflow against Neural Cleanse, STRIP, Fine-Pruning, Activation Clustering in dormant modeCapacity Overflow · ImageNet-100 and GTSRB (dormant mode)defense detection: NC 1.94, 1.37 on IN-100 (threshold 2.0); STRIP 3.5-9.0%Improved

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