Privacy, Security and Trust in Machine Learning Systems

1.

Subject title

Privacy, Security and Trust in Machine Learning Systems

Приватност, безбедност и доверба на системите за машинско учење

2.

Code

m23_w_055

3.

Study program

Cloud Computing, IT management, Bioinformatics, Security, Cryptography and Coding, Еducation with ICT, Eco-informatics, Inteligent Systems, Computer Science, Statistics and Data Analytics, Software for embedded systems, Software Engineering, Cloud Computing, IT management, Bioinformatics, Security, Cryptography and Coding, Statistics and Data Analytics, Software Engineering, Data science in computer science and engineering, Internet Technologies and cyber security,

4.

Organizer of the study program (unit, institute, department, division)

Faculty of Information Sciences and Computer Engineering

5.

Study cycle (first, second, third)

Втор циклус

6.

Academic year / semester

5 / Зимски

7. Number of ECTS credits

6.0

8.

Instructor

ворн. проф. д-р Ристе Стојанов проф. д-р Сашо Граматиков

9.

Prerequisites for enrollment

10.

Subject goals and competencies:

The goal of the course is introduction to the basic risks that may occur when applying machine learning in the software systems, the attacks that can be used to undermine their integrity, security and authority. the attacks that can be used to undermine their integrity, security and authority. In addition, strategies to protect systems from these most common attacks will be discussed. The second part of the course will address the challenges that exist with most systems that use machine learning, which is how to protect the privacy of the data used to train them. The final part of the course will focus on how to increase the confidence of machine learning systems by reviewing techniques for explaining their results.

11.

Subject content:

- Introduction to security dimensions, concepts, methods - ML perspecrive on security - ML solutions and their threats - Attacks Against ML - Choosing the right defense - ML privacy issues - Possible ML privacy solutions - ML confidence issues and solutions

12.

Learning methods:

Предавања поддржани со презентации преку слајдови, интерактивни предавања, вежби (користење на опрема и софтверски пакети), тимска работа, пример случаи, поканети гости предавачи, самостојна изработка и одбрана на проектна задача и семинарска работа, учење во електронско опкружување (форуми, консултации).

13.

Total available time fund

6.0 ECTS x 30 hours = 180 hours

14.

Time distribution

45 + 15 + 30 + 50 + 40 = 180 hours

15.

Forms of teaching activities

15.1.

Lectures - theoretical teaching

45 hours

15.2.

Exercises (laboratory, classroom), seminars, team work

15 hours

16.

Other forms of activities

16.1.

Project tasks

50 hours

16.2.

Independent tasks

30 hours

16.3.

Homework

40 hours

17.

Grading method

17.1.

Tests

45 points

17.2.

Seminar work / project (presentation: written and oral)

50 points

17.3.

Activities and learning

10 points

17.4.

Final exam

0 points

18.

Grading criteria (points / grade)

up to 50 points

5 (five) (F)

from 51 to 60 points

6 (six) (E)

from 61 to 70 points

7 (seven) (D)

from 71 to 80 points

8 (eight) (C)

from 81 to 90 points

9 (nine) (B)

from 91 to 100 points

10 (ten) (A)

19.

Condition for signature and taking final exam

реализирани активности

20.

Language of instruction

македонски и англиски

21.

Quality assurance method

механизам на интерна евалуација и анкети

22.

Literature

22.1.

Mandatory literature

22.2.

Additional literature