Courses – Fakultet tehničkih nauka u Novom Sadu

Subject: Mathematical basics of artificial intelligence (17.DOM60)

Native organizations units: No data

Type of studies	Title
Doctoral Academic Studies	Mathematics in Engineering (Godina: 1, Winter)
Doctoral Academic Studies	Information Systems Engineering (Godina: 1, Winter)
Doctoral Academic Studies	Mathematics in Engineering (Godina: 2, Winter)

General information:

Category	Scientific-professional
Scientific or art field	Teorijska i primenjena matematika
Interdisciplinary	Yes
ECTS	10

Educational goal:

Mastering the necessary theoretical knowledge in various fields of mathematics to fully understand and more easily master artificial intelligence techniques as well as selected application examples. The student is trained to use the appropriate software (Matlab fuzzy toolbox)

Educational outcome:

Acquired knowledge is the basis for understanding basic techniques of artificial intelligence and solving complex problems that require computer intelligence, and can not be solved using conventional mathematical approaches. The subject is of applicative nature. Thus, the scientific techniques are used in solving many practical applicative problems.

Course content:

Neural networks: feedforward (irreversible) neural networks; the noise of neural networks; spreads backward error; regularization in neural networks; Bayesian Networks; Deep-learning neural networks. Kernel methods: dual representations; core construction; radial function; maximum margin classifier; suport vector machine. Evolutionary methods: genetic algorithms; genetic programming; intelligence of the crowd; evolutionary strategies. Fuzzy systems: fuzzy sets; fuzzy logic; fuzzy relation; fuzzy decision-making.

Teaching methods:

Lectures. Consultations. Practical part of the material students work and lay in the computer laboratory by solving the mandatory tasks that are being evaluated. Programming is done in programming languages ??C and Matlab. Students can do optional tasks and they can earn additional points here. The agreed part of the material (which makes up the whole) is orally exhibited and delivered in written form as a seminar paper. Part of the materials that make up the logical whole can be taken as partial exams that are part of the exam. Partial exams are taken in written form. The oral part of the final exam is eliminatory.

Literature:

Authors	Title	Year	Publisher	Language
Kevin Gurney	An introduction to neural networks	1997	London and NewYork 1997 by UCL Press	English
M. P. Deisenroth, A. A. Faisal, C. S. Ong.	Mathematics for Machine Learning	2020	Cambridge University Press	English
Bishop, C.M.	Pattern Recognition and Machine Learning	2006	Springer, New York	English
S. Russell, P. Norvig	Artificial Intelligence: A Modern Approach	2007	Pearson Education Limited	English
Bezdek, J.C. et al.	Fuzzy models and algorithms for pattern recognition and image processing	1999	Kluwer Academic Publishers, Massachusetts	English

Knowledge evaluation:

Course activity	Pre-examination	Obligations	Number of points
Lecture attendance	Yes	Yes	5.00
Term paper	Yes	No	0.00
Oral part of the exam	No	Yes	15.00
Project task	Yes	Yes	40.00
Written part of the exam - tasks and theory	No	Yes	40.00
Homework	Yes	No	0.00