B4B36ZUI / BE4B36ZUI -- Introduction to Artificial Intelligence

The goal of this subject is to introduce the basics of artificial intelligence. We will cover the algorithms of informed and uninformed state space, problem solving methods, reinforcement learning, knowledge representation and (sequential) decision making under uncertainty. We will avoid most of classical machine learning, since that is the focus of other courses.

Lectures:

Viliam Lisý , Branislav Bošanský

Seminars:

Jakub Rada , Oleksii Shuhailo , Karolina Drabent , Jan Mrkos , Tomáš Holeček

Time tables: Czech English

BRUTE - homework submissions

Labs Notebooks (Google drive, use FEL Google Account)

• Getting the credit (“zápočet”)
• Passing the final exam
• The sum of the points for the final exam, midterm exam, and home works determines the final grade (50-59p. = E, …, 90-100p. = A).

The students can gain at most 30 points for homework assignments. In order to get the credit (zápočet), they have to submit each task for at least 5 points (before the penalisation for late submissions - penalisation does not prevent getting the credit) and gain at least 15 points for home works overall.

Penalisation for late submission:

• less than 24h after the deadline – losing 20% points
• more than 24h after the deadline – losing 100% points

Always work on your assignments individually. Plagiarism is being detected and it is not tolerated. If you have an objective reason for difficulties with finishing the assignment on time, contact us ASAP, please.

In the middle of the semester, there will be a test similar to the final exam for 15 points.

Example of the questions for midterm test:

• zui_example_questions.pdf
• zui_example_questions_2.pdf
• Example solved exercises

The final exam is for up to 55 points:

• the exam is in written form
• the students must gain at least 28 points to pass the exam
• the students with an overall sum of scores over 80 for the whole subject will have to pass also brief oral examination to defend the grade
• the topics of the questions for the exam are given by the slides, however, the slides are not meant to be the primary study materials and are not self-explanatory
• you can use a calculator during the exam (not a mobile phone, no other materials)
• the exam lasts 3 hours

Questions will be similar to the midterm test shown below, but they will cover all the content covered in the lectures and labs.

Exam dates:

• May 26, 9:00-12:00
• Jun 5, 9:00-12:00
• Jun 23, 13:00-16:00

Example exercises

Exam procedure:

1. You will be seated by the exam supervisors, with 2-seat gap between students and one row gap between rows - if the space in the room permits.
2. Exam supervisors will hand out the exams by calling out your name, checking your id and handing you the exam sheets.
3. Check your exam - your name should be printed on every page. Sign on the first page. Check that your exam has the correct number of questions.
4. Only start working on your exam when instructed by the exam supervisor.

Answering exam questions:

• If you think there are different ways to understand a question that can impact your answer → document your understanding and answer for that specifically.
• If you think there are assumptions missing you would need to answer a question, state them.
• If you are required to calculate an answer, try to provide the calculation steps towards your answer. You can potentially get full points for correct reasoning with numerical errors, but you get no points when you give only the incorrect result.
• Multiple choice questions can have 0, 1 or more correct answers. Specific questions may or may not provide additional instructions regarding correct number of answers.

Grading:

• No negative points from question, each question has 0 as minimum points.
• Multiple-choice questions are graded per examiners judgment. One bad answer can result in full points in the presence of textual explanation, but it can also result in 0 points if it demonstrates lack of understanding.
• We use best effort when grading exams, however, we make mistakes as well. After the exams are graded, we organize exam viewing, where you can check your answers and the points you received, and suggest corrections if you spot any errors (note, however, that this can result in point reduction as well).

• [AIMA] Russel, S. a Norvig, P.: Artificial Intelligence: A Modern Approach (2nd edition), Prentice Hall, 2003
  • příslušné kapitoly k dispozici na vyžádání
• [RLbook] Richard S. Sutton and Andrew G. Barto: Reinforcement Learning: An Introduction http://incompleteideas.net/book/the-book.html
• Ilustrace alpha-beta

More details on chapters and links to other resources are in the slides.