GARD cuts CLIP visual backdoor success below 1%
The post-hoc defense covers six visual backdoor attacks on two CLIP backbones and needs no extra inference compute.
Stories per day · 30 days
Relatedbackdoor2clip1computer-vision1contrastive-learning1multimodal1post-hoc1semantic-segmentation1
3 stories, newest first · one card per story
Stories tagged #model-defense
RePAIR repairs backdoored semantic segmentation models without retraining
The post-hoc, label-free method infers victim-target pairs from unlabeled images and corrects suspicious pixels.
ODPure: Backdoor Purification for Object Detection via Ensemble Corruption Consensus
ODPure purifies backdoor triggers in object detectors via corruption, reconstruction, and voting without discarding models.