- Prof.
**Tapio Elomaa** -
**Spring 2016**, Jan. 12 - Apr. 28 -
*Lectures*: Tue 12-14 TB219 and Thu 12-14 TB223 -
*Assignment*(Essay, Presentation, and Demonstration) is intructed by M.Sc. Juho Lauri -
*Course exam*: Wed May 4, 2016

- The presentation of Tue Apr. 19 and 26 will start at 12:15
- See free presentation times and topics here
- PRESENTATION TIMES: You will have a chance of giving the voluntary presentation concerning your essay topic on Tuesdays in April (Apr. 5, 12, 19, and 26) 10 - 12 in TB219. Please reserve a suitable time from Juho Lauri.
- Next lecture after Easter break on Thu Mar. 31.
- Assignment information updated on Feb. 16
- Assignment information

- This is one of Theoretical Computer Science courses at TUT. See the page for further information on courses and their relation to old courses.
- A course on programming
(e.g., TIE-02206
*Basic Course of Programming*) is an obligatory prerequisite. - Courses MAT-72006
*Advanced Algorithms and Data Structures*and OHJ-2306 Introduction to Theoretical Computer Science (MAT-73006 Theoretical Computer Science) help to follow the lectures and are, thus, recommended prerequisites. - The course can be incorporated into advanced studies of
*Computer Science*(30 cr) and*Mathematics*(30/50 cr) as well as international MSc major studies of*Mathematics*(30/50 cr). - The course is also suitable for postgraduate studies.

- The lectures are based on the textbook:
Russell & Norvig: Artificial Intelligence: A Modern Approach, Third ed., Pearson, 2010.

Week Dates Slide Numbers Chapters in the Course Book 1 Jan. 12 and 14 1-4, 24 - 39 13 Quantifying Uncertainty 2 Jan. 19 and 21 40 - 79 13, 14 Probabilistic Reasoning 3 Jan. 26 and 28 80 - 117 14, 16 Making Simple Decisions 4 Feb. 2 and 4 118 - 145 17 Making Complex Decisions 5 Feb. 9 and 11 146 - 167 18 Learning from Examples 6 Feb. 16 Juho's presentation on the assignment Feb. 18 168 - 183 18.4 - 18.5 7 Feb. 23 and 25 184 - 225 18.6 - 18.9 8 Mar. 8 and 10 226 - 253 20 Learning Probabilistic Models 9 Mar. 15 and 17 254 - 279 20, 21 Reinforcement Learning 10 Mar. 22 280 - 300 3 Solving Problems by Searching Mar. 31 301 - 315 3 Solving Problems by Searching 11 Apr. 5 and 7 316 - 349 4 Beyond Classical Search, 5 Adversarial Search 12 Apr. 12 and 14 350 - 379 7 Logical Agents 13 Apr. 19 and 21 380 - 410 8 First-Order Logic, 9 Inference in First-Order Logic 14 Apr. 26 Student presentations Apr. 28 411 - 434 9

- See free presentation times here
- PRESENTATION TIMES: You will have a chance of giving the
voluntary presentation concerning your essay topic on Tuesdays
in April (Apr. 5, 12, 19, and 26) 10 - 12 in TB219. Please
reserve a suitable time from Juho Lauri.
- The course assignment has three components:
- A hands-on empirical task,
- an essay, and
- a presentation on the essay topic.

- Tasks (1) and (3) are voluntary and do not necessarily need to be
taken on. Only (2) is required for passing the course. All subtasks
yield extra points for course grading. Tasks (1) and (2) both yield
max 5 points and the max for (3) is 3 points.
- The compulsory home work in the course is to write an essay on a freely chosen topic that is related to artificial intelligence in a reasonable way. You may choose a topic of your interest (after receiving an ok either from the lecturer Tapio Elomaa or Juho Lauri) or choose a topic from the list of potential candidates provided below. Appropriate length for the essay is 5-10 pages (using a reasonable font size and margins) and it should be written in English. The topic is not supposed to be new, but it should not only be based on the lectures. You may deal with a topic that is connected to the lectures, but you need to provide additional information from new sources. A suitable topic is a general overview on a research field, introduction to a specific topic, or something in between these extremes.
The essay must begin with an abstract of the contents. The main content is divided into sections and it ends with the list of references. You need to base your essay on at least three articles and they have to be referenced in the text. Own contributed text and citations have to be separated clearly. Whenever you make a claim which cannot be seen to be common knowledge among computer scientists, it has to be made clear from which source the claim comes from.

The intended audience for the essay is your fellow students knowledgeable in computer science and quick to learn, but no experts on the specific topic of your study. The essay should first introduce the background and required concepts for understanding details of its topic. The essay will be graded on the basis of general appearance, structure, language, and interestingness on the scale 0-5.

A short (15 min) presentation on the essay may be given on April. The presentation yields up to 3 extra points. Prepare to distribute some copies of your essay (or its short version) for the audience.

**Plagiarism is strictly prohibited. Do not base your essay on Wikipedia alone.**Deadline for the essay is May 8, 2016. Please deliver your essay to Juho Lauri preferably as pdf file.

- All tasks can be executed in groups of 1-3 students. Please try to
self-organize into groups. We can try to match people with similar
interests. Give us an e-mail if you cannot find collaborators on your own.
- Possible topics for the essay and presentation are given e.g. in
AAAI-16 Conference Keyword List.
You can modify the general topics by specifying a given key word or combining two key words into one topic. We particularly recommend categories:

Heuristic Search and Optimization

Machine Learning Methods

Search and Constraint SatisfactionYou may also propose a topic of your own. Please reserve a topic/confirm the suitability of your own topic with Juho.

The length of the essay is 5-10 pages. Groups with more members are expected to produce longer essays than students working on their own. The presentation length should be 15 minutes (not too many slides). Instructions for task (1) are given in http://math.tut.fi/~laurij/ai-assignment/

TIMETABLE

April presentations organized (time and place announced later)

May 4 course exam

May 8 deadline for tasks (1) and (2)

- We intend to cover the following chapters from the textbook:
Lecture week Topic 1-4 **IV Uncertain knowledge and reasoning**13 Quantifying Uncertainty 14 Probabilistic Reasoning 16 Making Simple Decisions 17 Making Complex Decisions 5-8 **V Learning**18 Learning from Examples 19 Knowledge in Learning 20 Learning Probabilistic Models 21 Reinforcement Learning 9-10 **II Problem-solving**3 Solving Problems by Searching 4 Beyond Classical Search 5 Adversarial Search 6 Constraint Satisfaction Problems 11-14 **III Knowledge, reasoning, and planning**7 Logical Agents 8 First-Order Logic 9 Inference in First-Order Logic 10 Classical Planning 11 Planning and Acting in the Real World

- Exam (May 4, 2016) maximum 30 points

- Stuart Russell & Peter Norvig: Artificial Intelligence: A Modern Approach, Third ed., Pearson, 2010

- AAAI - Association for the Advancement of Artificial Intelligence
- AiTopics
- Boston Dynamics
- IBM Watson
- VW Golf

Apr. 28, 2016 http://www.cs.tut.fi/~elomaa