Please wait ...
0% Complete
Home
/
15th International Conference on Computer and Knowledge Engineering
Towards Low-Overhead Mitigation of Trojan Bit-Flip Attacks on DNNs via Causal Inference
Authors :
Bahare Gholami
1
Mohsen Raji
2
1- دانشگاه شیراز
2- دانشگاه شیراز
Keywords :
Trojan Bit-Flip Attack،Lightweight Defense Mechanism،Causal Inference
Abstract :
Trojan bit-flip attacks pose a significant threat to the security of deep neural networks (DNNs), particularly in safety-critical applications. Existing defense mechanisms often incur substantial memory overhead, limiting their applicability in resource-constrained environments. Identifying the most vulnerable parts of a DNN is the key to designing efficient defense mechanisms against Trojan bit-flip attacks. In this paper, a low overhead defense mechanism for Trojan bit-flip attacks on DNNs is proposed. To identify the most influential layers within a DNN that can be the targets of attackers, the proposed approach leverages causal inference, a statistical and analytical framework for modeling the causal effects of interventions or changes in one variable (e.g., a Trojan bit-flip attack on the parameters of a DNN) on another (e.g., model predictions). By performing a layer-wise causal analysis, the method ranks layers according to their contribution to the model predictions and protects only those layers, rather than protecting all layers as in previous work. Evaluations on ResNet-32 using the CIFAR-10 dataset indicate that, relative to the state-of-the-art, the proposed method reduces the additional memory overhead by more than 3.5×. Specifically, it increases the base model size from 1.77 MB to 3.43 MB (an overhead of 1.66 MB), whereas the state-of-the art method increases it to 7.64 MB (an overhead of 5.87 MB). While preserving high model accuracy (89.64%) and achieving a similar reduction in attack success rate, these results highlight the effectiveness of the proposed causality-guided selective protection in improving DNN robustness with minimal overhead.
Papers List
List of archived papers
PersianILP: Construction and Evaluation of a Standard Persian Dataset for Inductive Link Prediction
Mohammad Rahimi - Afsaneh Fatemi - Ahmad Baraani
Analyzing the Impact of COVID-19 on Economy from the Perspective of User’s Reviews
Fatemeh Salmani - Hamed Vahdat-Nejad - Hamideh Hajiabadi
Characterizing Microsatellite Distribution Patterns Across Distinct Gene Categories in Human
Elahe Mehrazin - Mahmoud Naghibzadeh - Sara Jamali
Low-Cost and Hardware Efficient Implementation of Pooling Layers for Stochastic CNN Accelerators
Mobin Vaziri - Hadi Jahanirad
Enhancing Lighter Neural Network Performance with Layer-wise Knowledge Distillation and Selective Pixel Attention
Siavash Zaravashan - Sajjad Torabi - Hesam Zaravashan
Implementation of a Low-Overhead 2-Bit Parity-Preserving Reversible Vedic Multiplier for Quantum Architectures
Shekoofeh Moghimi - Negin Mashayekhi - Mohammad Reza Reshadinezhad
Improving performance of multi-label classification using ensemble of feature selection and outlier detection
Mohammad Ali Zarif - Javad Hamidzadeh
InfOnto: An ontology for fashion influencer marketing based on Instagram
Somaye Sultani - Mohsen Kahani
Robust Distributed Learning over Heterogeneous Adaptive Networks based on Federated BSP Model
Fatemeh Barani - MohammadHafez Yari - Abdorreza Savadi - Hadi Sadoghi Yazdi
Prediction of West Texas Intermediate Crude-oil Price Using Hybrid Attention-based Deep Neural Networks: A Comparative Study
Alireza Jahandoost - Mahboobeh Houshmand - Seyyed Abed Hosseini
more
Samin Hamayesh - Version 46.0.2