Nghia Dinh ndinh215

👷 What I work on

⬅ Shift Left

🔧 Early & Mid Stage | Analysis Tools

I am no longer abe to monitor this post , I have decided to move everything to my personal blog for better monitoring.

Please click here to access the full post

MNIST tutorial: the fast gradient sign method and adversarial training

This tutorial explains how to use CleverHans together with a TensorFlow model to craft adversarial examples, as well as make the model more robust to adversarial examples. We assume basic knowledge of TensorFlow.

Setup

First, make sure that you have TensorFlow

MNIST tutorial: crafting adversarial examples with the Jacobian-based saliency map attack

This tutorial explains how to use CleverHans together with a TensorFlow model to craft adversarial examples, using the Jacobian-based saliency map approach. This attack is described in details by the following paper. We assume basic knowledge of TensorFlow. If you need help getting CleverHans installed before getting started, you may find our MNIST tutorial on the fast gradient sign method to be useful.

	{
	"StartAt": "Add empty LastEvaluatedKey",
	"States": {
	"Add empty LastEvaluatedKey": {
	"Type": "Pass",
	"Next": "Scan",
	"Result": {
	"LastEvaluatedKey": null
	},
	"ResultPath": "$"

	package main

	import (
	"crypto/sha256"
	"encoding/json"
	"fmt"
	"strconv"
	"strings"
	"time"
	)

	#!/bin/bash

	### steps ####
	# verify the system has a cuda-capable gpu
	# download and install the nvidia cuda toolkit and cudnn
	# setup environmental variables
	# verify the installation
	###

	### to verify your gpu is cuda enable check

	from keras.models import Model
	from keras.layers.merge import concatenate
	from keras.layers import Conv2D , MaxPool2D , Input , GlobalAveragePooling2D ,AveragePooling2D, Dense , Dropout ,Activation, Flatten , BatchNormalization


	def conv_with_Batch_Normalisation(prev_layer , nbr_kernels , filter_Size , strides =(1,1) , padding = 'same'):
	x = Conv2D(filters=nbr_kernels, kernel_size = filter_Size, strides=strides , padding=padding)(prev_layer)
	x = BatchNormalization(axis=3)(x)
	x = Activation(activation='relu')(x)
	return x

	import cv2

	def add_gaussian_noise(X_imgs):
	gaussian_noise_imgs = []
	row, col, _ = X_imgs[0].shape
	# Gaussian distribution parameters
	mean = 0
	var = 0.1
	sigma = var ** 0.5

	Parallels Business Activation Key
	6YDASE-890T7S-E5YFS6-54F6YW-4780JY
	390349-032409-342340-932902-390693