ui培训怎么样_开发公司成本部职责岗位职责和流程_seo搜索排名优化是什么意思_网络营销方案设计毕业设计

之前已经针对CIFAR10训练了多种分类器：
Pytorch | 从零构建AlexNet对CIFAR10进行分类
Pytorch | 从零构建Vgg对CIFAR10进行分类
Pytorch | 从零构建GoogleNet对CIFAR10进行分类
Pytorch | 从零构建ResNet对CIFAR10进行分类
Pytorch | 从零构建MobileNet对CIFAR10进行分类
Pytorch | 从零构建EfficientNet对CIFAR10进行分类
Pytorch | 从零构建ParNet对CIFAR10进行分类

也实现了一些攻击算法：
Pytorch | 利用FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用BIM/I-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用MI-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用NI-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用PI-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用VMI-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用VNI-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用EMI-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用AI-FGTM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用I-FGSSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用SMI-FGRM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用VA-I-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用PC-I-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用IE-FGSM针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用GRA针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用GNP针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用MIG针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用DTA针对CIFAR10上的ResNet分类器进行对抗攻击
Pytorch | 利用PGN针对CIFAR10上的ResNet分类器进行对抗攻击

本篇文章我们使用Pytorch实现NCS对CIFAR10上的ResNet分类器进行攻击.

CIFAR数据集

CIFAR-10数据集是由加拿大高级研究所（CIFAR）收集整理的用于图像识别研究的常用数据集，基本信息如下：

• 数据规模：该数据集包含60,000张彩色图像，分为10个不同的类别，每个类别有6,000张图像。通常将其中50,000张作为训练集，用于模型的训练；10,000张作为测试集，用于评估模型的性能。
• 图像尺寸：所有图像的尺寸均为32×32像素，这相对较小的尺寸使得模型在处理该数据集时能够相对快速地进行训练和推理，但也增加了图像分类的难度。
• 类别内容：涵盖了飞机（plane）、汽车（car）、鸟（bird）、猫（cat）、鹿（deer）、狗（dog）、青蛙（frog）、马（horse）、船（ship）、卡车（truck）这10个不同的类别，这些类别都是现实世界中常见的物体，具有一定的代表性。

下面是一些示例样本：

NCS介绍

NCS（Neighborhood Conditional Sampling）攻击算法是一种用于增强对抗样本转移性的新型攻击方法，其核心思想是通过寻找具有高期望对抗损失和低标准偏差的对抗区域，以提高对抗样本在不同模型之间的转移性。以下是NCS攻击算法的详细介绍：

算法流程

1. 初始化：设置 $g_0=0$，$x_0^{adv}=x$，${\tilde{g}}_{-2,i}=0$，${\tilde{g}}_{-1,i}=0$，$\alpha =ϵ/T$。
2. 迭代更新：
   • 对于 $t=0$ 到 $T-1$：
     • 设置 $g=0$。
     • 对于 $i=0$ 到 $N-1$：
       • 随机采样一个示例 $x_{t,i}\in U_{\xi }(x^{adv})$。
       • 计算 $x_{t,i}^{\prime }=x_{t,i}-\gamma \cdot sign({\tilde{g}}_{t-2,i}-{\tilde{g}}_{t-1,i})$，并限制 $x_{t,i}^{\prime }\in B_{\gamma }(x_{t,i})$。
       • 计算关于 $x_{t,i}^{\prime }$ 的梯度 ${\tilde{g}}_{t,i}={\nabla }_{x_{t,i}^{\prime }}L(x_{t,i}^{\prime },y;F)$。
       • 累积更新梯度 $g=g+\frac{1}{N}\cdot {\tilde{g}}_{t,i}$。
     • 更新 $g_{t+1}=\mu \cdot g_t+\frac{g}{\Vert g{\Vert }_1}$。
     • 通过 $x_{t+1}^{adv}=x_t+\alpha \cdot sign(g_{t+1})$ 更新 $x_{t+1}^{adv}$， 并限制 $x_{t+1}^{adv}\in B_{\gamma }(x)$。
3. 返回：返回最终的对抗样本 $x^{adv}=x_T^{adv}$。

NCS代码实现

NCS算法实现

import torch
import torch.nn as nndef NCS(model, criterion, original_images, labels, epsilon, num_iterations=10, decay=1, N=20, xi=2 * (16 / 255), gamma=0.15 * (16 / 255), lambda_=1.6 / 255):"""NCS (Neighborhood Conditional Sampling) 攻击算法参数:- model: 要攻击的模型- criterion: 损失函数- original_images: 原始图像- labels: 原始图像的标签- epsilon: 最大扰动幅度- num_iterations: 迭代次数- decay: 动量衰减因子- N: 采样数量- xi: 邻域采样上限- gamma: 子区域上限- lambda_: 平衡系数"""alpha = epsilon / num_iterationsperturbed_images = original_images.clone().detach().requires_grad_(True)momentum = torch.zeros_like(original_images).detach().to(original_images.device)g_t_1 = torch.zeros_like(original_images).detach().to(original_images.device)g_t_2 = torch.zeros_like(original_images).detach().to(original_images.device)for t in range(num_iterations):accumulate_g = torch.zeros_like(original_images).detach().to(original_images.device)for _ in range(N):# 随机采样邻域内的点random_samples = original_images + (torch.rand_like(original_images) * 2 - 1) * xirandom_samples = random_samples.detach().requires_grad_(True)# 计算条件采样点x_i_prime = random_samples - gamma * torch.sign(g_t_2 - g_t_1)x_i_prime = torch.clamp(x_i_prime, random_samples - gamma, random_samples + gamma)x_i_prime = x_i_prime.detach().requires_grad_(True)outputs = model(x_i_prime)loss = criterion(outputs, labels)model.zero_grad()loss.backward()accumulate_g += x_i_prime.grad.data# 平均梯度g_t = decay * momentum + accumulate_g / torch.sum(torch.abs(accumulate_g), dim=(1, 2, 3), keepdim=True)momentum = g_t# 更新g_t_2和g_t_1g_t_2 = g_t_1g_t_1 = g_t# 计算对抗样本的更新方向sign_data_grad = g_t.sign()# 更新对抗样本perturbed_images = perturbed_images + alpha * sign_data_gradperturbed_images = torch.clamp(perturbed_images, original_images - epsilon, original_images + epsilon)perturbed_images = perturbed_images.detach().requires_grad_(True)return perturbed_images

攻击效果

为节省时间，这里取 N=1，实际实验中取 N=20

代码汇总

ncs.py

import torch
import torch.nn as nndef NCS(model, criterion, original_images, labels, epsilon, num_iterations=10, decay=1, N=20, xi=2 * (16 / 255), gamma=0.15 * (16 / 255), lambda_=1.6 / 255):"""NCS (Neighborhood Conditional Sampling) 攻击算法参数:- model: 要攻击的模型- criterion: 损失函数- original_images: 原始图像- labels: 原始图像的标签- epsilon: 最大扰动幅度- num_iterations: 迭代次数- decay: 动量衰减因子- N: 采样数量- xi: 邻域采样上限- gamma: 子区域上限- lambda_: 平衡系数"""alpha = epsilon / num_iterationsperturbed_images = original_images.clone().detach().requires_grad_(True)momentum = torch.zeros_like(original_images).detach().to(original_images.device)g_t_1 = torch.zeros_like(original_images).detach().to(original_images.device)g_t_2 = torch.zeros_like(original_images).detach().to(original_images.device)for t in range(num_iterations):accumulate_g = torch.zeros_like(original_images).detach().to(original_images.device)for _ in range(N):# 随机采样邻域内的点random_samples = original_images + (torch.rand_like(original_images) * 2 - 1) * xirandom_samples = random_samples.detach().requires_grad_(True)# 计算条件采样点x_i_prime = random_samples - gamma * torch.sign(g_t_2 - g_t_1)x_i_prime = torch.clamp(x_i_prime, random_samples - gamma, random_samples + gamma)x_i_prime = x_i_prime.detach().requires_grad_(True)outputs = model(x_i_prime)loss = criterion(outputs, labels)model.zero_grad()loss.backward()accumulate_g += x_i_prime.grad.data# 平均梯度g_t = decay * momentum + accumulate_g / torch.sum(torch.abs(accumulate_g), dim=(1, 2, 3), keepdim=True)momentum = g_t# 更新g_t_2和g_t_1g_t_2 = g_t_1g_t_1 = g_t# 计算对抗样本的更新方向sign_data_grad = g_t.sign()# 更新对抗样本perturbed_images = perturbed_images + alpha * sign_data_gradperturbed_images = torch.clamp(perturbed_images, original_images - epsilon, original_images + epsilon)perturbed_images = perturbed_images.detach().requires_grad_(True)return perturbed_images

train.py

import torch
import torch.nn as nn
import torchvision
import torchvision.transforms as transforms
from models import ResNet18# 数据预处理
transform_train = transforms.Compose([transforms.RandomCrop(32, padding=4),transforms.RandomHorizontalFlip(),transforms.ToTensor(),transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))
])transform_test = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010))
])# 加载Cifar10训练集和测试集
trainset = torchvision.datasets.CIFAR10(root='./data', train=True, download=False, transform=transform_train)
trainloader = torch.utils.data.DataLoader(trainset, batch_size=128, shuffle=True, num_workers=2)testset = torchvision.datasets.CIFAR10(root='./data', train=False, download=False, transform=transform_test)
testloader = torch.utils.data.DataLoader(testset, batch_size=100, shuffle=False, num_workers=2)# 定义设备（GPU或CPU）
device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")# 初始化模型
model = ResNet18(num_classes=10)
model.to(device)# 定义损失函数和优化器
criterion = nn.CrossEntropyLoss()
optimizer = torch.optim.Adam(model.parameters(), lr=0.01)if __name__ == "__main__":# 训练模型for epoch in range(10):  # 可以根据实际情况调整训练轮数running_loss = 0.0for i, data in enumerate(trainloader, 0):inputs, labels = data[0].to(device), data[1].to(device)optimizer.zero_grad()outputs = model(inputs)loss = criterion(outputs, labels)loss.backward()optimizer.step()running_loss += loss.item()if i % 100 == 99:print(f'Epoch {epoch + 1}, Batch {i + 1}: Loss = {running_loss / 100}')running_loss = 0.0torch.save(model.state_dict(), f'weights/epoch_{epoch + 1}.pth')print('Finished Training')

advtest.py

import torch
import torch.nn as nn
import torchvision
import torchvision.transforms as transforms
from models import *
from attacks import *
import ssl
import os
from PIL import Image
import matplotlib.pyplot as pltssl._create_default_https_context = ssl._create_unverified_context# 定义数据预处理操作
transform = transforms.Compose([transforms.ToTensor(),transforms.Normalize((0.491, 0.482, 0.446), (0.247, 0.243, 0.261))])# 加载CIFAR10测试集
testset = torchvision.datasets.CIFAR10(root='./data', train=False,download=False, transform=transform)
testloader = torch.utils.data.DataLoader(testset, batch_size=128,shuffle=False, num_workers=2)# 定义设备（GPU优先，若可用）
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")model = ResNet18(num_classes=10).to(device)criterion = nn.CrossEntropyLoss()# 加载模型权重
weights_path = "weights/epoch_10.pth"
model.load_state_dict(torch.load(weights_path, map_location=device))if __name__ == "__main__":# 在测试集上进行FGSM攻击并评估准确率model.eval()  # 设置为评估模式correct = 0total = 0epsilon = 16 / 255  # 可以调整扰动强度for data in testloader:original_images, labels = data[0].to(device), data[1].to(device)original_images.requires_grad = Trueattack_name = 'NCS'if attack_name == 'FGSM':perturbed_images = FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'BIM':perturbed_images = BIM(model, criterion, original_images, labels, epsilon)elif attack_name == 'MI-FGSM':perturbed_images = MI_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'NI-FGSM':perturbed_images = NI_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'PI-FGSM':perturbed_images = PI_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'VMI-FGSM':perturbed_images = VMI_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'VNI-FGSM':perturbed_images = VNI_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'EMI-FGSM':perturbed_images = EMI_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'AI-FGTM':perturbed_images = AI_FGTM(model, criterion, original_images, labels, epsilon)elif attack_name == 'I-FGSSM':perturbed_images = I_FGSSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'SMI-FGRM':perturbed_images = SMI_FGRM(model, criterion, original_images, labels, epsilon)elif attack_name == 'VA-I-FGSM':perturbed_images = VA_I_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'PC-I-FGSM':perturbed_images = PC_I_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'IE-FGSM':perturbed_images = IE_FGSM(model, criterion, original_images, labels, epsilon)elif attack_name == 'GRA':perturbed_images = GRA(model, criterion, original_images, labels, epsilon)elif attack_name == 'GNP':perturbed_images = GNP(model, criterion, original_images, labels, epsilon)elif attack_name == 'MIG':perturbed_images = MIG(model, original_images, labels, epsilon)elif attack_name == 'DTA':perturbed_images = DTA(model, criterion, original_images, labels, epsilon)elif attack_name == 'PGN':perturbed_images = PGN(model, criterion, original_images, labels, epsilon)elif attack_name == 'NCS':perturbed_images = NCS(model, criterion, original_images, labels, epsilon)perturbed_outputs = model(perturbed_images)_, predicted = torch.max(perturbed_outputs.data, 1)total += labels.size(0)correct += (predicted == labels).sum().item()accuracy = 100 * correct / total# Attack Success RateASR = 100 - accuracyprint(f'Load ResNet Model Weight from {weights_path}')print(f'epsilon: {epsilon:.4f}')print(f'ASR of {attack_name} : {ASR :.2f}%')