0% Complete
Home
/
12th International Conference on Computer and Knowledge Engineering
GAP: Fault tolerance Improvement of Convolutional Neural Networks through GAN-aided Pruning
Authors :
Pouya Hosseinzadeh
1
Yasser Sedaghat
2
Ahad Harati
3
1- Department of Computer Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
2- Department of Computer Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
3- Department of Computer Engineering, Ferdowsi University of Mashhad, Mashhad, Iran
Keywords :
CNN،fault tolerance،auxiliary pruning task،generalization،GAN،N-version programming،ensemble learning،bit-flip fault
Abstract :
Functionality and accuracy degradation imposed by Bit-Flip faults (BF) on the CNNs' weights makes it vital first to evaluate neural networks' behavior in the presence of faults in the design phase and then make use of these models, especially in safety-critical applications like autonomous vehicles. In the following paper, we propose a novel approach to improve the fault tolerance of convolutional neural networks through pruning their parameters with regard to their gradients instead of the more commonly used norm values. In particular, the gradient of the network's parameters on an auxiliary classification task based on Generative Adversarial Network (GAN) is exploited to identify parameters of lower generalization. Due to the inherent efficiency that all pruning techniques introduce into the model, it is practical to use the ensembled of pruned models. In this paper, we evaluate the functionality of the naïve ensemble versus the ensemble of three differently pruned models. This idea comes from the N-version Programming (NVP) concept of dependability literature. To assess the resiliency of our model, we conduct comprehensive random bit-flip fault injection experiments. Comparing the result of our approach to ordinary pruning techniques, we show that the classification accuracy is improved by 20% in the presence of a single fault in the network. Making an NVP ensemble of three VGG-16 GAN-aided pruned networks, we make a 10% further accuracy improvement in the presence of a single random fault in each base-learner for the CIFAR-10 dataset.
Papers List
List of archived papers
Soccer Video Event Detection Using Metric Learning
Ali Karimi - Ramin Toosi - Mohammad Ali Akhaee
A parallel CNN-BiGRU network for short-term load forecasting in demand-side management
Arghavan Irankhah - Sahar Rezazadeh Saatlou - Mohammad Hossein Yaghmaee - Sara Ershadi-Nasab - Mohammad Alishahi
An Interactive Approach for Query-based Multi-Document Scientific Text Summarization
Mohammadsadra Nejati - Azadeh Mohebi - Abbas Ahmadi
A Stacking Ensemble Framework for Ransomware Detection on the Bitcoin Blockchain Using Transaction Graph Analytics
Mohammad Mobin Teymourpour - Parsa Hedayatnia - Mohammad Allahbakhsh - Haleh Amintoosi
Capturing Local and Global Features in Medical Images by Using Ensemble CNN-Transformer
Javad Mirzapour Kaleybar - Hooman Saadat - Hooman Khaloo
Automated software design using Machine Learning With Natural Language Processing
Fahimeh Khedmatkon - Seyed Mohammad Hossein Hasheminejad - Jaleh Shoshtarian Malak
Cloud Service Composition Using Genetic Algorithm and Particle Swarm Optimization
Javad Dogani - Farshad Khunjush
A Robust Network for Embedded Traffic Sign Recognation.
Omid Nejati Manzari - Shahriar Baradaran Shokouhi
Design and Simulation of a Low PDP Full Adder by Combining Majority Function and TGDI Technique in CNTFET Technology
Mahsa Mohammadi
DevRanker: An Effective Approach to Rank Developers for Bug Report Assignment
Mohammad Reza Kardoost - Mohammad Reza Moosavi - Reza Akbari
more
Samin Hamayesh - Version 42.7.0