# 经济代写|博弈论代写Game Theory代考|Parallel Scanning

#### Doug I. Jones

Lorem ipsum dolor sit amet, cons the all tetur adiscing elit

couryes-lab™ 为您的留学生涯保驾护航 在代写博弈论Game Theory方面已经树立了自己的口碑, 保证靠谱, 高质且原创的统计Statistics代写服务。我们的专家在代写博弈论Game Theory代写方面经验极为丰富，各种代写博弈论Game Theory相关的作业也就用不着说。

• Statistical Inference 统计推断
• Statistical Computing 统计计算
• (Generalized) Linear Models 广义线性模型
• Statistical Machine Learning 统计机器学习
• Longitudinal Data Analysis 纵向数据分析
• Foundations of Data Science 数据科学基础
couryes™为您提供可以保分的包课服务

## 经济代写|博弈论代写Game Theory代考|Parallel Scanning

Currently, most Internet scans (e.g., scans in the Censys database) are performed separately and independently across different ports. In other words, the entire IPv4 address is sweeped multiple times, each time sending probes to all IP addresses on a certain port. This allows different ports to be scanned independently, possibly at different times, thereby reducing the amount of traffic sent to networks/hosts. In this scenario, our method can only use the location and AS properties of the targeted IP addresses for predicting the responses of hosts, as depicted in Figure 21.2a. In this diagram, the geolocation (GL) and AS features are fed to each trained model in order to produce the prediction $\hat{y}_i^k$ of the true label $y_i^k$ for sample $i$ and port $k$, i.e., the estimated likelihood that IP address $i$ will respond to probes on port $k$. These predictions are then fed to the scanner, which will decide whether to scan different IP/port pairs depending on the prediction of the model. In this study, we make decisions by thresholding $\hat{y}_i^k$; if $\hat{y}_i^k<t_r^k$ the scanner refrains from sending the probe. Note that $t_r^k$ (specific to port $k$ ) is the threshold for reaching a target true positive rate $r$.
While this approach uses a minimal amount of information for prediction, applying machine learning to parallel scans is fairly straightforward, since the predictions of trained models can simply be translated into blacklists that can be fed to network scanners for refraining from sending probes to certain IP/port pairs. Moreover, due to the crude granularity of geolocation and AS features, we do not need to perform predictions for every IP address, but only for IP blocks in which all IP addresses share the same features, therefore reducing the computational overhead of our approach. We will discuss this point in more detail in Section 21.6.4.

## 经济代写|博弈论代写Game Theory代考|Sequential Scanning

In contrast to parallel scanning, one can also design a scanner to scan different ports in a sequential manner. In this setting, we can take advantage of the responses of previously scanned ports for predicting the remaining labels. Cross-protocol dependencies have been observed by Bano et al. (2018), but were not directly used for bootstrapping network scans. This is due to the fact that cross-protocol correlations by themselves are not sufficient for predicting other port labels, as we will further discuss in Section 21.6.1. However, we show that when combined with prescan features, i.e., location and AS properties, cross-protocol information can help improve the efficiency of sequential scans, as compared to parallel scans.

Assume $x_i^k, k \in{1, \ldots, M}$ to denote the feature vector resulting from probing IP $i$ on port $k$, and $x_i^0$ to denote prescan features. Then for sequential scanning, the classifier for port $k$ is trained using $\left{x_i^l, l<k\right}$ as features, i.e., GL/AS features, as well as ports scanned earlier in the sequence. Note that for parallel scans in Section 21.4.1 we are only performing predictions using $x_i^0$ as features. We evaluate and compare this approach to parallel scanning in Section 21.5, resulting in more bandwidth savings. Figure 21.2b depicts the process used for sequential scans. Similar to parallel scanning, each model in this figure is generating a prediction $\hat{y}_i^k$ for an IP/port pair, which is then fed to the scanner for thresholding. The features resulting from each scan (i.e., the postscan features in Section 21.3.2) are then appended to the model’s input features and used for all subsequent models. This allows models toward the end of the sequence to make predictions based on a richer feature set, which can result in more bandwidth savings for their corresponding scans.

# 博弈论代考

## 有限元方法代写

tatistics-lab作为专业的留学生服务机构，多年来已为美国、英国、加拿大、澳洲等留学热门地的学生提供专业的学术服务，包括但不限于Essay代写，Assignment代写，Dissertation代写，Report代写，小组作业代写，Proposal代写，Paper代写，Presentation代写，计算机作业代写，论文修改和润色，网课代做，exam代考等等。写作范围涵盖高中，本科，研究生等海外留学全阶段，辐射金融，经济学，会计学，审计学，管理学等全球99%专业科目。写作团队既有专业英语母语作者，也有海外名校硕博留学生，每位写作老师都拥有过硬的语言能力，专业的学科背景和学术写作经验。我们承诺100%原创，100%专业，100%准时，100%满意。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。

Days
Hours
Minutes
Seconds

# 15% OFF

## On All Tickets

Don’t hesitate and buy tickets today – All tickets are at a special price until 15.08.2021. Hope to see you there :)