Task

黑箱攻击(Blackbox Attack)

采用CIFAR-10数据集经助教筛选，黑箱攻击的核心是：如果你有目标网络的训练数据，你可以训练一个proxy network，用这个proxy network生成attacked objects。

Baseline

• Simple baseline (acc <= 0.70)
  • Hints: FGSM
  • Expected running time: 1.5 mins on T4
• Medium baseline (acc <= 0.50)
  • Hints: Ensemble Attack + ensemble random few model + IFGSM
  • Expected running time: 5 mins on T4
• Strong baseline (acc <= 0.25)
  • Hints: Ensemble Attack + ensemble many models + MIFGSM
  • Expected running time: 5.5 mins on T4
• Boss baseline (acc <= 0.10)
  • Hints: Ensemble Attack + ensemble many models + DIM-MIFGSM
  • Expected running time: 8 mins on T4

主要就是以下这几个方法，具体的理论写在ipynb中了

Implement non-targeted adversarial attack method

a. FGSM
b. I-FGSM
c. MI-FGSM

Increase attack transferability by Diverse input (DIM)

Attack more than one proxy model - Ensemble attack

FGSM (Fast Gradient Sign Method (FGSM)

$$
x^{{adv}的目标是J(x}{real},y)最大\

x^{adv} = x^{real} + ϵ ⋅ sign(\frac{\partial{J(x^{real},y)}}{\partial{x}0})
$$

I-FGSM(Iterative Fast Gradient Sign Method)

$$\begin{array}{rl}& {\mathbf{x}}_0^{adv}={\mathbf{x}}^{real}\\ & \mathrm{for}\mathrm{t}=1\mathrm{to}num\_iter:\\ & \boxed{{\mathbf{x}}_{t+1}^{adv}=\mathbf{x}{\ast }_t^{adv}+\alpha \cdot \mathrm{sign}({\nabla }_{\ast }\mathbf{x}J({\mathbf{x}}_t^{adv},y))}\\ & \mathrm{clip}{\mathbf{x}}_t^{adv}\end{array}$$

其中，$\alpha$是step size

MI-FGSM(Momentum Iterative Fast Gradient Sign Method)

使用momentum来稳定更新方向，并摆脱糟糕的局部最大值
KaTeX parse error: Got function '\boldsymbol' with no arguments as subscript at position 112: …t+\frac{\nabla_\̲b̲o̲l̲d̲s̲y̲m̲b̲o̲l̲{x}J(\boldsymbo…
其中$g$是momentum

M-DI2-FGSM(Diverse Input Momentum Iterative Fast Gradient Sign Method)

$$\begin{gathered} T(X_n^{adv};p)=\{\begin{array}{ll}T(X_n^{adv})& \text{ with probability }p\\ X{\ast }_n^{adv}& \text{ with probability }1-p\end{array} \\ \text{e.g. DIM + MI-FGSM} \\ {g}_{\ast }n+1=\mu \cdot g_n+\frac{{\nabla }_{X}L(T(X_n^{adv};p),y^{\mathrm{true}};\theta )}{||{\nabla }_{X}L(T(X_n^{adv};p),y^{\mathrm{true}};\theta )|{|}_1} \end{gathered}$$

这里的L可以用CrossEntropyLoss求解

Report

fgsm attack

# original image
path = f'dog/dog2.png'
im = Image.open(f'./data/{path}')
logit = model(transform(im).unsqueeze(0).to(device))[0]
predict = logit.argmax(-1).item()
prob = logit.softmax(-1)[predict].item()
plt.title(f'benign: dog2.png\n{classes[predict]}: {prob:.2%}')
plt.axis('off')
plt.imshow(np.array(im))
plt.tight_layout()
plt.show()

# adversarial image
adv_im = Image.open(f'./fgsm/{path}')
logit = model(transform(adv_im).unsqueeze(0).to(device))[0]
predict = logit.argmax(-1).item()
prob = logit.softmax(-1)[predict].item()
plt.title(f'adversarial: dog2.png\n{classes[predict]}: {prob:.2%}')
plt.axis('off')
plt.imshow(np.array(adv_im))
plt.tight_layout()
plt.show()

Jepg Compression

import imgaug.augmenters as iaa

# pre-process image
x = transforms.ToTensor()(adv_im)*255
x = x.permute(1, 2, 0).numpy()
x = x.astype(np.uint8)

# TODO: use "imgaug" package to perform JPEG compression (compression rate = 70)
compressed_x = iaa.JpegCompression(compression=70)(image=x)

logit = model(transform(compressed_x).unsqueeze(0).to(device))[0]
predict = logit.argmax(-1).item()
prob = logit.softmax(-1)[predict].item()
plt.title(f'JPEG adversarial: dog2.png\n{classes[predict]}: {prob:.2%}')
plt.axis('off')


plt.imshow(compressed_x)
plt.tight_layout()
plt.show()

Code Link

代码在Github