Improved Gradient based Adversarial Attacks for Quantized Networks
- URL: http://arxiv.org/abs/2003.13511v2
- Date: Wed, 29 Dec 2021 17:22:40 GMT
- Title: Improved Gradient based Adversarial Attacks for Quantized Networks
- Authors: Kartik Gupta, Thalaiyasingam Ajanthan
- Abstract summary: We show that quantized networks suffer from gradient vanishing issues and show a fake sense of robustness.
By attributing gradient vanishing to poor forward-backward signal propagation in the trained network, we introduce a simple temperature scaling approach to mitigate this issue.
- Score: 15.686134908061995
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Neural network quantization has become increasingly popular due to efficient
memory consumption and faster computation resulting from bitwise operations on
the quantized networks. Even though they exhibit excellent generalization
capabilities, their robustness properties are not well-understood. In this
work, we systematically study the robustness of quantized networks against
gradient based adversarial attacks and demonstrate that these quantized models
suffer from gradient vanishing issues and show a fake sense of robustness. By
attributing gradient vanishing to poor forward-backward signal propagation in
the trained network, we introduce a simple temperature scaling approach to
mitigate this issue while preserving the decision boundary. Despite being a
simple modification to existing gradient based adversarial attacks, experiments
on multiple image classification datasets with multiple network architectures
demonstrate that our temperature scaled attacks obtain near-perfect success
rate on quantized networks while outperforming original attacks on
adversarially trained models as well as floating-point networks. Code is
available at https://github.com/kartikgupta-at-anu/attack-bnn.
Related papers
- Quantization-aware Interval Bound Propagation for Training Certifiably
Robust Quantized Neural Networks [58.195261590442406]
We study the problem of training and certifying adversarially robust quantized neural networks (QNNs)
Recent work has shown that floating-point neural networks that have been verified to be robust can become vulnerable to adversarial attacks after quantization.
We present quantization-aware interval bound propagation (QA-IBP), a novel method for training robust QNNs.
arXiv Detail & Related papers (2022-11-29T13:32:38Z) - Dynamics-aware Adversarial Attack of Adaptive Neural Networks [75.50214601278455]
We investigate the dynamics-aware adversarial attack problem of adaptive neural networks.
We propose a Leaded Gradient Method (LGM) and show the significant effects of the lagged gradient.
Our LGM achieves impressive adversarial attack performance compared with the dynamic-unaware attack methods.
arXiv Detail & Related papers (2022-10-15T01:32:08Z) - DenseShift: Towards Accurate and Efficient Low-Bit Power-of-Two
Quantization [27.231327287238102]
We propose the DenseShift network, which significantly improves the accuracy of Shift networks.
Our experiments on various computer vision and speech tasks demonstrate that DenseShift outperforms existing low-bit multiplication-free networks.
arXiv Detail & Related papers (2022-08-20T15:17:40Z) - Dynamics-aware Adversarial Attack of 3D Sparse Convolution Network [75.1236305913734]
We investigate the dynamics-aware adversarial attack problem in deep neural networks.
Most existing adversarial attack algorithms are designed under a basic assumption -- the network architecture is fixed throughout the attack process.
We propose a Leaded Gradient Method (LGM) and show the significant effects of the lagged gradient.
arXiv Detail & Related papers (2021-12-17T10:53:35Z) - Meta Adversarial Perturbations [66.43754467275967]
We show the existence of a meta adversarial perturbation (MAP)
MAP causes natural images to be misclassified with high probability after being updated through only a one-step gradient ascent update.
We show that these perturbations are not only image-agnostic, but also model-agnostic, as a single perturbation generalizes well across unseen data points and different neural network architectures.
arXiv Detail & Related papers (2021-11-19T16:01:45Z) - A Layer-wise Adversarial-aware Quantization Optimization for Improving
Robustness [4.794745827538956]
We find that adversarially-trained neural networks are more vulnerable to quantization loss than plain models.
We propose a layer-wise adversarial-aware quantization method, using the Lipschitz constant to choose the best quantization parameter settings for a neural network.
Experiment results show that our method can effectively and efficiently improve the robustness of quantized adversarially-trained neural networks.
arXiv Detail & Related papers (2021-10-23T22:11:30Z) - On the Adversarial Robustness of Quantized Neural Networks [2.0625936401496237]
It is unclear how model compression techniques may affect the robustness of AI algorithms against adversarial attacks.
This paper explores the effect of quantization, one of the most common compression techniques, on the adversarial robustness of neural networks.
arXiv Detail & Related papers (2021-05-01T11:46:35Z) - Implicit Under-Parameterization Inhibits Data-Efficient Deep
Reinforcement Learning [97.28695683236981]
More gradient updates decrease the expressivity of the current value network.
We demonstrate this phenomenon on Atari and Gym benchmarks, in both offline and online RL settings.
arXiv Detail & Related papers (2020-10-27T17:55:16Z) - Boosting Gradient for White-Box Adversarial Attacks [60.422511092730026]
We propose a universal adversarial example generation method, called ADV-ReLU, to enhance the performance of gradient based white-box attack algorithms.
Our approach calculates the gradient of the loss function versus network input, maps the values to scores, and selects a part of them to update the misleading gradients.
arXiv Detail & Related papers (2020-10-21T02:13:26Z) - Mixed-Precision Quantized Neural Network with Progressively Decreasing
Bitwidth For Image Classification and Object Detection [21.48875255723581]
A mixed-precision quantized neural network with progressively ecreasing bitwidth is proposed to improve the trade-off between accuracy and compression.
Experiments on typical network architectures and benchmark datasets demonstrate that the proposed method could achieve better or comparable results.
arXiv Detail & Related papers (2019-12-29T14:11:33Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.