Course Description

The course will cover the entire data science process from data preparation, feature engineering and selection, exploratory data analysis, data visualization, machine learning, model evaluation & optimization and finally, implementing at scale. Participants will learn how to read data in a variety of common formats, and then write scripts to analyze and visualize the data in meaningful ways. The course is specifically designed to provide sophisticated training in data science as applied to cybersecurity-related challenges & scenarios.

Learning Modules

Note - this content is not finalized and may be subject to change prior to release.

• Introduction to Data Science

  This module introduces the course, as well as key concepts of data science and their application to security.

  • Define Python

  • Introduce the data science process

  • Discuss case studies of the application of data science and machine learning to security

  • Introduce the Python data science ecosystem to include Jupyter Notebook and various Python modules

  • Introduce the Griffon VM

• Exploratory Data Analysis in One Dimension

  This module introduces the concept of vectorized computing, and how to create, manipulate and summarize one dimensional data using the Pandas module in Python. Additionally, we will cover basic statistical concepts.

  • Create a series object

  • Explore and summarize data within a Series

  • Understand and generate Tukey 5 number summaries, as well as calculate other common statistical measures

• Exploratory Data Analysis in Two Dimensions

  This module builds on the concepts taught in module 1, and introduces the students to the DataFrame: a two dimensional, vectorized data structure. This module also covers how to directly ingest security data into a Pandas DataFrame.

  • Create a DataFrame using the various read_ functions

  • Import various security data into dataframes

  • Flatten complex nested data

  • Join and merge data sets

  • Calculate aggregate statistics

• Data Visualization

  Data visualization is a powerful technique to have in your analytic toolkit. This module will cover the theory of data visualization as well as the actual process and coding of creating effective visualizations.

  • Understand techniques about how to make effective data visualizations

  • Be familiar with various modules in Python to create visualizations

  • Be able to create both static and interactive visualizations

• Machine Learning - Feature Engineering

  The Machine Learning modules walk the student through the machine learning process from beginning to end, starting with feature engineering and selection, model creation and ultimately evaluating and improving model performance.

  • Complete walkthrough of the machine learning process

• Supervised Machine Learning

  This module introduces students to the concepts behind classifiers and various classification algorithms. Students will also learn to evaluate classifier performance.

  • Understand the functioning of a classifier and various classification algorithms including Support Vector Machines, Decision Trees, and Random Forest

  • Be able to create models using various classification algorithms in Python using Scikit-Learn

  • Be able to evaluate the performance of a model and tune its hyperparameters.

• Unsupervised Learning

  This module introduces students to unsupervised learning and how to apply it to security problems. The module mainly focuses on clustering and its uses and limitations.

  • Understand various distance measurement functions

  • Understand the concepts behind K-Means and DBSCAN

  • Be able to create clusterers and evaluate their performance

• Anomaly Detection

  This module covers various techniques that can be used to detect anomalies in security data. This module will introduce the students to various approaches to anomaly detection, including forecasting, unsupervised machine learning and other statistical techniques.

  • Understand different approaches for detecting anomalies in security data

  • Be able to implement common unsupervised machine learning algorithms such as one-class support vector machines and isolation forests to detect anomalous data

  • Be able to perform anomaly detection using statistical metrics such as the Grubbs’ test

  • Understand the challenges associated with using Machine Learning to detect anomalies

• Hunting with Data Science

  Deep learning is one of the most exciting and new areas in machine learning. In this module, we will give the students a conceptual overview of deep learning and its application to security problems. Given the complexity of this topic, this module is intended to be a more conceptual introduction, not an in-depth technical course.

  • Understand the basic concepts behind deep neural nets (DNN)

  • Become familiar the Python deep learning ecosystem

  • Understand the concepts behind more advanced NN such as convolutional NNs

• Deep Learning

  This module covers the topics of deep learning, neural networks, convolutional neural networks, recurrent neural networks, and their applications in cybersecurity. It also discusses the use of deep learning tools such as TensorFlow and Keras.

  • Gain understanding of deep learning concepts including neural networks, CNNs, and RNNs.

  • Learn how to apply deep learning techniques to cybersecurity tasks.

  • Acquire practical skills in using TensorFlow, Keras, and Word2Vec for implementing deep learning models and processing cybersecurity data.

• Hacking Machine Learning Models

  One of the most interesting topics in security data science is the possibility that machine learning models can be hacked. This module covers the real-world implications of hacking machine learning models, and the techniques used to hack.

  • Understand the risks of deploying ML models

  • Understand the techniques used to hack ML models

  • Understand ways of minimizing risk to a production model

• Course Assessement

  Complete a course assessment to earn a certificate of completion!