Course Title: Artificial Intelligence – VSŠ (Bologna) – 2nd Year

Lecturer:

Prof. Dr. Igor Kononenko
Assistants: Dr. Petar Vračar, Dr. Tome Eftimov

Course Objective:

To introduce machine learning and artificial intelligence methods and develop the ability to apply them practically.

Student Obligations:

  • Timely submission of homework assignments with positive evaluation (online quizzes and progress reports)
  • Timely submission of both research seminar papers with positive evaluation
  • Written and oral exam

Grading

Exercises

Students solve online quizzes posted on the e-learning platform by the deadlines, which are graded. Each student must achieve at least 50% in the online quizzes to receive a grade for the exercises. The grade for the exercises is based on the combined grade from the two research seminar papers. Each seminar paper must be submitted on time and graded positively. A condition for a positive grade in exercises is also achieving at least half of the points in homework assignments. Only timely submitted homework assignments are considered. The exercise grade is valid only for the current year. If a student does not pass the exam in the current year, they must resubmit the homework and research seminar papers in the following academic year.

Exam

The exam consists of a written and possibly an oral part. The condition for taking the exam is a positive evaluation of the exercises. During the written exam, students are allowed one A4 sheet handwritten with a regular pencil (so it can be erased) and signed with a pen, including their name, surname, and student number (photocopies and printed pages are not allowed). This sheet is handed in along with the written exam.

Final Grade

The final grade is composed of the exercise grade (50%) and the overall exam grade (50%). The student must achieve at least half of the possible points in each part, both in the exercises and in the written exam.

Approximate Content:

  1. Introduction to machine learning and overview of machine learning methods
  2. What is intelligence, what is learning, and the human-machine relationship
  3. Basic principles of machine learning
  4. Decision trees and regression trees
  5. Basic machine learning methods
  6. Attribute evaluation, data preprocessing
  7. Evaluation of learning, ensembles, artificial neural networks
  8. Search spaces: uninformed, local, heuristic
  9. Searching using the MINIMAX principle, game playing
  10. Evolutionary computation and genetic algorithms
  11. Intelligent agents and robots
  12. Reinforcement learning

Primary Literature:

I. Kononenko and M. Robnik Šikonja: Intelligent Systems. FE and FRI Publishing, Ljubljana, 2010.

Additional Literature:

  • I. Kononenko and M. Kukar: Machine Learning and Data Mining: Introduction to Principles and Algorithms. Horwood Publishing, 2007.
  • G. F. Luger: Artificial Intelligence: Structures and Strategies for Complex Problem Solving (6th ed.). Addison-Wesley, Pearson Education, 2009.
  • D. L. Poole, A. K. Mackworth: Artificial Intelligence: Foundations of Computational Agents. Cambridge University Press, 2010.
  • S. J. Russell, P. Norvig: Artificial Intelligence: A Modern Approach (3rd ed.). Prentice Hall, Pearson Education, 2010.

EXAMPLES OF RESEARCH SEMINAR PAPERS

1. Research Seminar Paper: Machine Learning

Students are divided into two groups during each cycle of laboratory exercises. Each group selects a leader. The leader manages the entire project of analyzing two databases using machine learning methods. Each database must be approached in several ways:

  • Data visualization
  • Attribute evaluation: detection of random and redundant attributes, selection of important attribute subsets
  • As a classification problem (by discretizing the dependent variable into 2 intervals with a given threshold)
  • As a regression problem

Each group will have subcontractors (subgroups) for 4 tasks. The leader coordinates the group’s work by dividing it into subgroups, assigning tasks to each subgroup, coordinating the results from each subgroup to ensure they complement each other, for example:

  • Visualization and selection of important attribute subsets
  • Comparison of classification and regression trees, etc.

Each group prepares a report on the analysis of both databases, and each student writes up to 4 pages about their work. The leader’s main report (also up to 4 pages) summarizes the entire group’s work, and appendices to the report include the reports from all contributors.

Groups defend their research seminar paper at the scheduled time during exercises, with the leader presenting the group’s work. Afterward, the assistant asks questions to each group member individually. Each group also prepares the “best” classifier and “best” regressor for both problems. To test them, the group receives a specified number of new test examples for both problems. The “estimated” accuracy is compared with the actual accuracy on the new test examples.

2. Research Seminar Paper: Search Spaces