Improved Gradient based Adversarial Attacks for Quantized Networks
This work addresses robustness issues in quantized networks, which are critical for efficient AI deployment, but it is incremental as it modifies existing gradient-based attacks.
The authors tackled the problem of gradient vanishing in quantized neural networks, which leads to a false sense of robustness against adversarial attacks, and introduced a temperature scaling method that achieves near-perfect success rates on quantized networks while also improving performance on adversarially trained and floating-point models.
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.