我们知道模型的结构，就可以进行攻击了，这个是白盒攻击，但是大部分时候，模型的结构我们无从得知，可以进行黑盒攻击。黑盒攻击就是攻击代理模型，比方说对vggnet的攻击在ResNet上同样管用。下面是一些数据，来证明黑盒攻击有用。（表格里的数字表示正确率）
三、FGSM攻击
论文地址：https://arxiv.org/abs/1412.6572
求解一个y =
四、天池人脸识别对抗
1、比赛地址：https://tianchi.aliyun.com/competition/entrance/231745/information
2、比赛评分规则：
为了保证扰动后人脸的视觉效果，本次比赛限制单个像素的扰动在[-25.5, 25.5]区间内，对于扰动超出该范围的提交结果，我们会在后台强制把图像扰动截断至[-25.5, 25.5]区间(使用numpy.clip函数)。所以请参赛选手控制提交的对抗样本与原图在单像素上的差异。
对每个生成的对抗样本，后台会采用模型对该样本进行预测，并根据识别结果计算相应的扰动量，具体计算公式如下：
其中 M表示后台模型预测结果， y表示样本 I的真实标签。如果防御算法对样本识别正确，此次攻击不成功，扰动量直接置为上限44.1673。该上限可由约束的最大扰动25.5计算得出。如果攻击成功，计算对抗样本 I^a 和原始样本I的 L2 距离，作为得分，得分越小越好。
一句话描述规则就是改动越小，攻击成功率越高，成绩越好。
五、DeepLearning4j进行FGSM攻击
1、解决dl4j对input求梯度问题
我们攻击的代理模型同样是VggFace（为什么一直都是选vggface，确实dl4j只有vggface，哎，也没有其他选择），ComputationGraph中梯度反向传播完成Gradient对象就被回收了 ，这样设计的目的就是为了节省内存。在MultiLayerNetwork中可以通过org.deeplearning4j.nn.multilayer.MultiLayerNetwork#calculateGradients求Loss对input的偏导数，但我们怎么拿的ComputationGraph中Loss对input的偏导数呢？不着急，我们在源码中找答案，我们细看源码org.deeplearning4j.nn.graph.ComputationGraph#calcBackpropGradients中反向传播做了什么？
try (MemoryWorkspace wsWorkingMem = workspaceMgr.notifyScopeEntered(ArrayType.BP_WORKING_MEM)) {
pair = current.doBackward(truncatedBPTT, workspaceMgr);
epsilons = pair.getSecond();
//Validate workspace location for the activation gradients:
//validateArrayWorkspaces(LayerWorkspaceMgr mgr, INDArray array, ArrayType arrayType, String vertexName, boolean isInputVertex, String op){
for (INDArray epsilon : epsilons) {
if (epsilon != null) {
//May be null for EmbeddingLayer, etc
validateArrayWorkspaces(workspaceMgr, epsilon, ArrayType.ACTIVATION_GRAD, vertexName, false, "Backprop");
}
}
} 跟进org.deeplearning4j.nn.graph.vertex.GraphVertex#doBackward方法
public Pair doBackward(boolean tbptt, LayerWorkspaceMgr workspaceMgr) {
if (!canDoBackward()) {
if(inputs == null || inputs[0] == null){
throw new IllegalStateException("Cannot do backward pass: inputs not set. Layer: \"" + vertexName
+ "\" (idx " + vertexIndex + "), numInputs: " + getNumInputArrays());
} else {
throw new IllegalStateException("Cannot do backward pass: all epsilons not set. Layer \"" + vertexName
+ "\" (idx " + vertexIndex + "), numInputs :" + getNumInputArrays() + "; numOutputs: "
+ getNumOutputConnections());
}
}
//Edge case: output layer - never did forward pass hence layer.setInput was never called...
if(!setLayerInput){
applyPreprocessorAndSetInput(workspaceMgr);
}
Pair pair;
if (tbptt && layer instanceof RecurrentLayer) {
//Truncated BPTT for recurrent layers
pair = ((RecurrentLayer) layer).tbpttBackpropGradient(epsilon,
graph.getConfiguration().getTbpttBackLength(), workspaceMgr);
} else {
//Normal backprop
pair = layer.backpropGradient(epsilon, workspaceMgr); //epsTotal may be null for OutputLayers
}
if (layerPreProcessor != null) {
INDArray eps = pair.getSecond();
eps = layerPreProcessor.backprop(eps, graph.batchSize(), workspaceMgr);
pair.setSecond(eps);
}
//Layers always have single activations input -> always have single epsilon output during backprop
return new Pair<>(pair.getFirst(), new INDArray[] {pair.getSecond()});
}
里面有个org.deeplearning4j.nn.conf.InputPreProcessor#backprop将梯度回传给InputPreProcessor处理。于是我们就有思路了，我们只需要给第一层卷积层设置一个InputPreProcessor即可获取回传的梯度，注意LayerVertex的InputPreProcessor是final修饰的，那么怎么设置InputPreProcessor呢？这个难不倒Javaer，反射。
public class LayerVertex extends BaseGraphVertex {
private Layer layer;
private final InputPreProcessor layerPreProcessor;
private boolean setLayerInput; 接下来先实现一个InputPreProcessor，把回传的梯度放在一个static变量里
public class Preprocessor implements InputPreProcessor {
private static final long serialVersionUID = 1L;
public static INDArray epsilon;
@Override
public INDArray preProcess(INDArray input, int miniBatchSize, LayerWorkspaceMgr workspaceMgr) {
return workspaceMgr.dup(ArrayType.ACTIVATIONS, input);
}
@Override
public InputType getOutputType(InputType inputType) {
return inputType;
}
@Override
public Pair feedForwardMaskArray(INDArray maskArray, MaskState currentMaskState,
int minibatchSize) {
return null;
}
@Override
public INDArray backprop(INDArray output, int miniBatchSize, LayerWorkspaceMgr workspaceMgr) {
epsilon = output.detach();
return workspaceMgr.dup(ArrayType.ACTIVATION_GRAD, output);
}
@Override
public InputPreProcessor clone() {
// TODO Auto-generated method stub
return null;
}
} 接下来先dl4j transfer learning API加载vggface的模型，去掉全连接层，加上CnnLossLayer作为output，这里Loss函数用的COSINE_PROXIMITY（尝试过多种方法之后，发现cosine距离效果最好），然后反射给第一层卷积层加上InputPreProcessor，反射时调用Field的setAccessible(true)方法，开放private属性的访问权限（当然这是迫不得已的方法），请看下面代码。
ComputationGraph pretrained = (ComputationGraph) VGG16.builder().build().initPretrained(PretrainedType.VGGFACE);
System.out.println(pretrained.summary());
FineTuneConfiguration fineTuneConf = new FineTuneConfiguration.Builder().updater(new Sgd(0)).seed(123).build();
ComputationGraph vgg16Transfer = new TransferLearning.GraphBuilder(pretrained)
.fineTuneConfiguration(fineTuneConf).removeVertexAndConnections("flatten")
.removeVertexAndConnections("fc6").removeVertexAndConnections("fc7").removeVertexAndConnections("fc8")
.addLayer("out", new CnnLossLayer.Builder(LossFunctions.LossFunction.COSINE_PROXIMITY)
.activation(Activation.IDENTITY).build(), "pool5")
.setOutputs("out").build();
LayerVertex conv1_1 = (LayerVertex) vgg16Transfer.getVertex("conv1_1");
Class clz = conv1_1.getClass();
Field nameField = clz.getDeclaredField("layerPreProcessor");
nameField.setAccessible(true);
nameField.set(conv1_1, new Preprocessor());
System.out.println(vgg16Transfer.summary()); 到此为止，Loss对input的偏导数就可以通过Preprocessor.epsilon获取到了，这个问题解决了，就可以进行攻击了。最终代理模型的结构如下：
================================================================================================
VertexName (VertexType) nIn,nOut TotalParams ParamsShape Vertex Inputs
================================================================================================
input_1 (InputVertex) -,- - - -
conv1_1 (ConvolutionLayer) 3,64 1,792 W:{64,3,3,3}, b:{1,64} [input_1]
conv1_2 (ConvolutionLayer) 64,64 36,928 W:{64,64,3,3}, b:{1,64} [conv1_1]
pool1 (SubsamplingLayer) -,- 0 - [conv1_2]
conv2_1 (ConvolutionLayer) 64,128 73,856 W:{128,64,3,3}, b:{1,128} [pool1]
conv2_2 (ConvolutionLayer) 128,128 147,584 W:{128,128,3,3}, b:{1,128} [conv2_1]
pool2 (SubsamplingLayer) -,- 0 - [conv2_2]
conv3_1 (ConvolutionLayer) 128,256 295,168 W:{256,128,3,3}, b:{1,256} [pool2]
conv3_2 (ConvolutionLayer) 256,256 590,080 W:{256,256,3,3}, b:{1,256} [conv3_1]
conv3_3 (ConvolutionLayer) 256,256 590,080 W:{256,256,3,3}, b:{1,256} [conv3_2]
pool3 (SubsamplingLayer) -,- 0 - [conv3_3]
conv4_1 (ConvolutionLayer) 256,512 1,180,160 W:{512,256,3,3}, b:{1,512} [pool3]
conv4_2 (ConvolutionLayer) 512,512 2,359,808 W:{512,512,3,3}, b:{1,512} [conv4_1]
conv4_3 (ConvolutionLayer) 512,512 2,359,808 W:{512,512,3,3}, b:{1,512} [conv4_2]
pool4 (SubsamplingLayer) -,- 0 - [conv4_3]
conv5_1 (ConvolutionLayer) 512,512 2,359,808 W:{512,512,3,3}, b:{1,512} [pool4]
conv5_2 (ConvolutionLayer) 512,512 2,359,808 W:{512,512,3,3}, b:{1,512} [conv5_1]
conv5_3 (ConvolutionLayer) 512,512 2,359,808 W:{512,512,3,3}, b:{1,512} [conv5_2]
pool5 (SubsamplingLayer) -,- 0 - [conv5_3]
out (CnnLossLayer) -,- 0 - [pool5]
------------------------------------------------------------------------------------------------
Total Parameters: 14,714,688
Trainable Parameters: 14,714,688
Frozen Parameters: 0
================================================================================================ 2、生成Label张量
下面把需要攻击的目标图片下载下来，我放在D盘了，目标图片如下。
下面用vggFace读取所有图片，把图片转化为张量，我们只需要获取最后一个池化层的输出就可以了。请看下面代码
NativeImageLoader loader = new NativeImageLoader(224, 224, 3, new ResizeImageTransform(224, 224));
File file = new File("D:/securityAI_round1_images/images");
ImageLoader imageLoader = new ImageLoader(112, 112, 3);
List list = new ArrayList<>();
for (File f : file.listFiles()) {
list.add(f);
}
Map labelMap = new HashMap<>();
for (int i = 0; i < list.size(); i++) {
vgg16Transfer.clear();
INDArray image = loader.asMatrix(list.get(i)).div(255);
Map map = vgg16Transfer.feedForward(image, false);
labelMap.put(i, map.get("pool5"));
} 3、无目标攻击
用第2步获取的张量作为label，用gradient ascent方法找到COSINE_PROXIMITY的极大值，COSINE_PROXIMITY的实现里将consine加了负号，所以是求cosine的最小值。换句话讲就是找到一张改动最小，且最不像自己的图片。代码如下
for (int i = 0; i < list.size(); i++) {
vgg16Transfer.clear();
INDArray oldImage = loader.asMatrix(list.get(i)).div(255);
INDArray newImage = oldImage;
NesterovsUpdater nesterovsUpdater = new NesterovsUpdater(0.9, 0.01, new long[] { 1, 3, 224, 224 });
for (int m = 0; m < 2; m++) {
vgg16Transfer.setInputs(newImage);
vgg16Transfer.setLabels(
labelMap.get(i).add(Nd4j.rand(new long[] { 1, 512, 7, 7 }, new NormalDistribution(0, 0.07))));
vgg16Transfer.computeGradientAndScore();
INDArray epsilon = Preprocessor.epsilon;
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(0, 30), NDArrayIndex.all())// 去掉30行
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(0, 75),
NDArrayIndex.interval(0, 60))// 额头
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(0, 75),
NDArrayIndex.interval(164, 224))// 额头
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(30, 45),
NDArrayIndex.interval(0, 60))// 额头
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(30, 45),
NDArrayIndex.interval(164, 224))// 额头
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(0, 75),
NDArrayIndex.interval(72, 80))// 额头
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(0, 75),
NDArrayIndex.interval(135, 152))// 额头
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(75, 115),
NDArrayIndex.interval(0, 40))// 眼睛
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(75, 115),
NDArrayIndex.interval(184, 224))// 眼睛
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(115, 165),
NDArrayIndex.interval(0, 40))// 脸
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(115, 165),
NDArrayIndex.interval(179, 224))// 脸
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(165, 195),
NDArrayIndex.interval(0, 50))// 嘴巴
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(165, 195),
NDArrayIndex.interval(174, 224))// 嘴巴
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(195, 224),
NDArrayIndex.interval(0, 70))// 下巴
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(195, 224),
NDArrayIndex.interval(154, 224))// 下巴
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(195, 224),
NDArrayIndex.interval(75, 97))// 下巴
.assign(0);
epsilon.get(NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.interval(195, 224),
NDArrayIndex.interval(127, 149))// 下巴
.assign(0);
epsilon = Transforms.sign(epsilon);
nesterovsUpdater.applyUpdater(epsilon);
INDArray preUpdate = newImage.add(epsilon);
INDArray delta = oldImage.sub(preUpdate);
INDArray tooLarge = delta.dup();// 因为减的太多了
BooleanIndexing.replaceWhere(tooLarge, 0, Conditions.absLessThanOrEqual(max));
BooleanIndexing.replaceWhere(tooLarge, 0, Conditions.lessThan(0));
tooLarge.subi(max);
BooleanIndexing.replaceWhere(tooLarge, 0, Conditions.lessThan(0));
INDArray tooSmall = delta.dup();// 因为加的太多了
BooleanIndexing.replaceWhere(tooSmall, 0, Conditions.absLessThanOrEqual(max));
BooleanIndexing.replaceWhere(tooSmall, 0, Conditions.greaterThan(0));
tooSmall.addi(max);
BooleanIndexing.replaceWhere(tooSmall, 0, Conditions.greaterThan(0));
INDArray bias = tooLarge.add(tooSmall);
newImage = preUpdate.add(bias);
vgg16Transfer.clear();
System.out.println(vgg16Transfer.score());
}
System.out.println("平均偏差：" + (max - Transforms.abs(oldImage.sub(newImage)).meanNumber().doubleValue()));
System.out.println("===========================");
newImage = newImage.mul(255);
BooleanIndexing.replaceWhere(newImage, 0, Conditions.lessThan(0.0));
BooleanIndexing.replaceWhere(newImage, 255, Conditions.greaterThan(255));
BufferedImage bufferedImage = new BufferedImage(224, 224, BufferedImage.TYPE_INT_RGB);
imageLoader.toBufferedImageRGB(newImage.get(new INDArrayIndex[] { NDArrayIndex.point(0), NDArrayIndex.all(),
NDArrayIndex.all(), NDArrayIndex.all() }), bufferedImage);
ImageIO.write(bufferedImage, "jpg", new File("D:/preImage/" + list.get(i).getName()));
INDArray oldSmallLoader = originalLoad.asMatrix(list.get(i)).div(255);
INDArray smallImage = smallLoader.asMatrix(new File("D:/preImage/" + list.get(i).getName()));
System.out.println("原始平均偏差："
+ (max - Transforms.abs(oldSmallLoader.sub(smallImage.div(255))).meanNumber().doubleValue()));
System.out.println("===========================");
BufferedImage smallBufferedImage = new BufferedImage(112, 112, BufferedImage.TYPE_INT_RGB);
imageLoader.toBufferedImageRGB(smallImage.get(new INDArrayIndex[] { NDArrayIndex.point(0),
NDArrayIndex.all(), NDArrayIndex.all(), NDArrayIndex.all() }), smallBufferedImage);
ImageIO.write(smallBufferedImage, "jpg", new File("D:/images/" + list.get(i).getName()));
} 说明：
（1）、目标图片为112*112的图片，经过尝试，先拓宽为224*224图片后再攻击，效果会提升。（推测dl4j vggface是用224*224的人脸进行训练，所以做了这个尝试）。
（2）、在label中随机加入微小的噪声，可以提升效果，这段代码就是这个原因labelMap.get(i).add(Nd4j.rand(new long[] { 1, 512, 7, 7 }, new NormalDistribution(0, 0.07)))。
（3）、在gradient ascent的过程中，同样加入动量作为更新因素效果会好一点，代码中的NesterovsUpdater就是实现这个功能。
（4）、代码中BooleanIndexing.replaceWhere是为了将图像的改动量限定在一个范围上，实际上是一个clip操作，却写了这么多代码，这一点确实不如Python方便。
（5）、为了进步提升分数，在更新图片时，最对嘴巴、眼睛、鼻子处进行变更，也能提升分数。
（6）、目标Loss Function为COSINE_PROXIMITY，之前也尝试过MSE和MAE效果不佳。
4、最终生成的攻击样本
关于如何用DL4J对人脸识别模型进行攻击问题的解答就分享到这里了，希望以上内容可以对大家有一定的帮助，如果你还有很多疑惑没有解开，可以关注创新互联行业资讯频道了解更多相关知识。
分享题目：如何用DL4J对人脸识别模型进行攻击
网站URL：http://www.cdkjz.cn/article/jsssdc.html
之后，让猫被识别为了狗，这就是攻击的过程。
