Introduction to artificial intelligence

Completed course Programming 1.

Introduction to Artificial Intelligence. From the Turing Test to contemporary definitions of artificial intelligence. Basic concepts: rational agent and its interaction with the environment, systems of larger agents, uncertainty, and machine learning. Ethical aspects of AI. Automatic decision-making by intelligent agents and the consequences of decision-making. Data bias in machine learning.
Heuristic search. Problem solving with search, search space. Basic search algorithms, depth- and breadth-first, and best-first search. Comparison of the computational complexity of the basic algorithms. Heuristic functions, algorithm A, acceptable heuristic functions, and optimality of A.
Heuristic search for playing zero-sum games. Min-Max Algorithm. Min-Max search trees and pruning search trees with the Alpha-Beta Algorithm.
Constraint satisfaction. Solving problems with constraints. Constraint formalization and propositional logic. Formulation of general SAT and 3-SAT problems. Algorithms for solving the general 3-SAT problem.
Dealing with uncertainty and Bayesian networks. The topology of Bayesian networks and its interpretation in terms of a condensed representation of the distribution of random variables. Calculating the (conditional) probabilities of events, i.e., Bayesian network nodes. Algorithm for efficient computation of event probabilities in Bayesian networks.
Machine learning. Basic problems of supervised machine learning, overfitting to training data, and curse of dimensionality. Loss functions and performance measurement of models. Procedures for evaluating the performance of models: train and test data. Bias-variance decomposition of error.
Overview of machine learning algorithms. Naive Bayes. Decision trees and tree ensembles. Linear models for regression and classification. Neural networks.

• Stuart Russell in Peter Norvig (2020) Artificial Intelligence: A Modern Approach. Četrta izdaja. Pearson. Spletna stran https://aima.cs.berkeley.edu. Izbrana poglavja.
• Gareth James, Daniela Witten, Trevor Hastie, Rob Tibshirani (2021) An Introduction to Statistical Learning. Druga izdaja. Springer. Prosto dostopna na https://www.statlearning.com. Izbrana poglavja.

Students get familiar with the basic concepts and algorithms of artificial intelligence and acquire practical knowledge for implementing simple algorithms and using existing implementations of complex artificial intelligence algorithms.

Knowledge and understanding: Knowledge and understanding of the basic artificial intelligence algorithms and practical problems in which they can be meaningfully used. Understanding the theoretical background of how algorithms work.
Application: Formulation of practical problems into a form that allows solving with artificial intelligence algorithms. Design and implementation of solutions using artificial intelligence algorithms.
Reflection: Critical analysis and assessment of the usefulness of AI algorithms for solving a chosen problem. Analysis of the ethical implications of the use of artificial intelligence.
Transferable skills – not tied to just one subject: Algorithmic thinking, the ability to identify computer problems and choose suitable algorithms.

Lectures, exercises, homeworks, seminar/group work, consultations.

Homeworks and written exam
Theoretical exam

Ljupčo Todorovski:
BRENCE, Jure, DŽEROSKI, Sašo, TODOROVSKI, Ljupčo. Dimensionally-consistent equation discovery through probabilistic attribute grammars. [Print ed.]. 2022, vol. 632, str. 742-756. ISSN 0020-0255. DOI: 10.1016/j.ins.2023.03.073.
LUKŠIČ, Žiga, TANEVSKI, Jovan, DŽEROSKI, Sašo, TODOROVSKI, Ljupčo. Meta-model framework for surrogate-basedparameter estimation in dynamical systems. IEEE access. 2019, vol. 7, str. 181829 -181841. ISSN 2169-3536. DOI: 10.1109/ACCESS.2019.2959846.
GRAU LEGUIA, Marc, LEVNAJIĆ, Zoran, TODOROVSKI, Ljupčo, ŽENKO, Bernard. Reconstructing dynamical networks via feature ranking. Chaos. 2019, vol. 29, no. 9, str. 09310-1-093107-15. ISSN 1054-1500. DOI: 10.1063/1.5092170.