5 papers · 1 filter
CertMask: Certifiable Defense Against Adversarial Patches via Theoretically Optimal Mask Coverage
Xuntao Lyu, Ching-Chi Lin, Abdullah Al Arafat +3
Adversarial patch attacks inject localized perturbations into images to mislead deep vision models. These attacks can be physically deployed, posing serious risks to real-world app…
Unified Alignment Protocol: Making Sense of the Unlabeled Data in New Domains
Sabbir Ahmed, Mamshad Nayeem Rizve, Abdullah Al Arafat +4
Semi-Supervised Federated Learning (SSFL) is gaining popularity over conventional Federated Learning in many real-world applications. Due to the practical limitation of limited lab…
Fisher Information guided Purification against Backdoor Attacks
Nazmul Karim, Abdullah Al Arafat, Adnan Siraj Rakin +2
Studies on backdoor attacks in recent years suggest that an adversary can compromise the integrity of a deep neural network (DNN) by manipulating a small set of training samples. O…
Augmented Neural Fine-Tuning for Efficient Backdoor Purification
Nazmul Karim, Abdullah Al Arafat, Umar Khalid +2
Recent studies have revealed the vulnerability of deep neural networks (DNNs) to various backdoor attacks, where the behavior of DNNs can be compromised by utilizing certain types…
Efficient Backdoor Removal Through Natural Gradient Fine-tuning
Nazmul Karim, Abdullah Al Arafat, Umar Khalid +2
The success of a deep neural network (DNN) heavily relies on the details of the training scheme; e.g., training data, architectures, hyper-parameters, etc. Recent backdoor attacks…