Machine Learning in Cybersecurity

Machine Learning in Cyber Security

 

Course Description

Recent advances in Machine Learning has lead to near (or beyond) human-level performance in many tasks - autonomous driving, voice assistance, playing a variety of games. In terms of privacy and security, this is a double-edged sword. ML techniques can be used to efficiently detect and prevent attacks (e.g., intrusion detection). However, their deployment to many real-world sensitive systems (e.g., self-driving cars, the cloud) also makes them susceptible to numerous attacks, such as introducing imperceptible perturbations in inputs and forcing ML systems behave in unintended ways.

The course explores in-depth both of these sides to Machine Learning and Cyber Security. The content addresses the following areas:

1. ML overview
2. ML for improving security
3. Attacks on ML models
4. ML and Privacy

 

Class Time and Location

Winter Semester 2018/19
Lectures: Wednesday, 12:00 - 14:00  (starts at 12:15)
Exercises: Monday, 12:00 - 14:00  (starts at 12:15)
Location: E91, 0.05

Instructor: Mario Fritz
Teaching Assistants: Tribhuvanesh Orekondy, Hossein Hajipour, Kathrin Grosse

 

Schedule and Syllabus

Event Type		Date	Description	Course Materials
Lecture		October 17	Course Overview	Slides
Tutorial		October 22	Primer on Python and Jupyter	Slides Notebook | Collab-hosted
Lecture		October 24	Machine Learning Overview	Slides
Tutorial		October 29	Hands-on ML with Python and Scikit-learn	Notebook | Collab-hosted
Project 1				Notebook | Collab-hosted Due on: 11-Nov-2018 (23:59)
Lecture		October 31	Deep Learning Overview	Slides
Tutorial		November 5	Applying Deep Learning	Notebook numpy | Notebook Keras
Lecture		November 7	Anomaly Detection	Slides
Lecture		November 14	Evasion Attacks	Slides
Tutorial		November 19	Project 1: Summary and Discussion	Slides
Project 2				Notebook Due on: 9-Dec-2018 (23:59)
Lecture		November 21	Membership Inference	Slides
Tutorial		November 26	Q&A (TA office hours)
Lecture		November 28	Malware Analysis	Slides
Project 3				Handout Due on: 3-Feb-2019 (23:59)
Lecture		December 12	Model Stealing & Watermarking	Slides
Tutorial		December 17	Project 2: Summary and Discussion	Slides
Lecture		December 19	Poisoning	Slides
Lecture		January 9	Network Traffic Analysis	Slides
			Using the GPU Machines	Slides
Lecture		January 16	Generative Adversarial Networks	Slides
Lecture		January 23	Differential Privacy	Slides
Lecture		February 6	Variational Auto-Encoders	Slides

 

Prerequisites

• Programming: Basic programming skills and familiarity with Python. All assignments will be in Python.
• Machine Learning: Prior knowledge in ML is helpful.
• Linear Algebra, Probability, Statistics and Calculus: Introductory level.

 

Course Discussions

For course-related discussion (or to find team-mates for projects), you can optionally use the following google group:
https://groups.google.com/d/forum/ml-cysec-2018

 

Project Honor Code

We encourage discussing ideas and concepts with other students to help you learn and better understand the course content. However, the work you submit and present must be original and demonstrate your effort in solving the presented problems. We will not tolerate blatantly using existing solutions (such as from the internet), improper collaboration (e.g., sharing code or experimental data between groups) and plagiarism. If the honor code is not met, no points will be awarded.

Project Presentation Schedule

Place E91, 0.01 (Small lecture hall, ground floor, right corner.), except the meeting on Feb. 8th, which takes place in the top floor rotunda at MPI.

Feb 8th	Group	Feb 11th	Group	Feb 13th	Group
10:00	Watchers on the Wall	14:00	The A-Team	14:00	Team 1
10:20	DKK	14:20	Winter is coming	14:20	Rage against ML
		14:40	Pandas	14:40	Gradient Ninjas
		15:00	Fastest Neighbors	15:00	Overestimators
		15:20	- - -	15:20	Security Monks
		15:40	- - -	15:40	Fast 3