|
This course has two parts. The first part includes an introduction to the
basic machine learning concepts and algorithms, which will also provide the basis for the second
part of the course. The second part covers selected recent topics in machine learning.
In particular, the course will cover the following main topics:
|
Part 1: |
| |
- Concept learning
- Decision trees
- Artificial neural networks
- Evaluating hypotheses
- Bayesian learning
- Instance based learning
- Genetic algorithms
- Hidden Markov models
- Reinforcement learning
|
|
Part 2: |
| |
- Evaluation methods for machine learning
- Cost-sensitive learning
- Ensemble learning
- Learning with multiple sources
- Active learning
- Transfer learning
- Kernel methods
- Learning with graphs
- Structured output learning
|
Homework (30%)
Paper presentation (15%)
Paper reports (10%)
Paper discussion (10%)
Project (35%)
Homework assignments will be posted on this web site. Assignments will
have some programming and non-programming parts and you are expected to work individually for the assignments.
Assignments are expected to be turned in at class time on the due date. For the late assignments,
you will be given a total of three grace days (whole or partial) for the whole semester. Once these late days have been
exhausted, no late assignments will be accepted. As an example, if you submit your 1st assignment 29 hours late, you will
have used two late days and have only one day left. If you then submit your 2nd assignment 5 hours late, you will have
used your remaining late day. If you submit your 3rd assignment 1 minute late, this assignment will not be accepted.
In the second part of the course, we will read and discuss research papers on the selected
recent topics of machine learning. In this part, you will present a paper in the class. The papers are
selected and assigned to the students by the instructor. If a paper is relatively longer and/or harder to
understand, it may be assigned to a group of two students.
In the paper assignments, we will try to match the topics with students' interest. Therefore,
you are asked to email your preferences on the topics, which are listed as the topics of the second part, to the
instructor by October 14th, 2011, if you have any preferences. If there are some topics for which several students
show interest, the related topics are assigned to the students on first-come first-served basis. If there are some
topics for which no one shows interest, the related papers are assigned to the remaining students. Thus, you are advised
to give more than one preference. Additionally, you are welcome to make suggestions for paper selection. However,
there is no guarantee that your suggested paper is selected and assigned to you. Please email your suggestions to
the instructor, if you have any.
Students will have 20-25 minutes to present their paper. We will have a discussion period of
10-15 minutes after the presentation. Students presenting their papers in a group of two will have 30-35 minutes
to complete their presentations. We will have 15-20 minutes for the discussion. You are supposed to submit
your presentations to the instructor and these presentations will also be graded. As a presenter, you should read
and understand the paper thoroughly and prepare a presentation that explains the main parts of the paper and
criticizes it. Your presentation should include
- The problem domain
- The main motivation of the paper
- The main contribution(s) of the paper
- The method proposed by the paper
- The very brief explanation of the experiments and their results
- The parts that you are in favor of
- The parts that you are against of
Moreover, all students are expected to read the papers presented by the other students
and participate the discussion session. As a participant of a discussion session, you are expected to ask
relevant questions to the presenter(s) and answer the questions related to the paper. For that, it is
important for you to read and understand the paper before coming to the presentation. Additionally, at the
beginning of each presentation, you are asked to submit a short report about the paper. This report should include
- The main motivation of the paper
- The main contribution(s) of the paper
- One part that you are in favor of
- One part that you are against of
This report should be very brief. Each item listed above should be explained in a short paragraph
that is typically composed of 2-4 sentences. Shorter answers, of course if they are correct, will be given extra credits.
You will conduct a research project individually or in a group of two.
Your project should contain a relevant part from the topics, which are listed as the topics of the
second part of the course. Additionally, you are required to use proper techniques to evaluate
your work (you may want to use some techniques from what we will mention under the topic of
"Evaluation Methods for Machine Learning").
You are required to write a proposal, a progress report, and a final report.
Your proposal should provide a short description of the problem that you want to work on. You are expected
to define your own problems. However, the instructor may approve the proposal or may want to change your problem.
Therefore, it is highly recommended to talk to the instructor on what you want to work on beforehand.
Its due date is November 14th, 2011.
The progress report should give the details of the problem and the related literature.
It should include the main steps of the methodology that you use in implementing what you proposed. It should
also give the experimental setup and evaluation techniques that you plan to use. Its due date is December 16th, 2011.
The final report should include the final steps of the methodology, your experiments, their
results, and your interpretations. It should also contain thorough literature search. Its due date is January 6th, 2012.
This course follows the Bilkent University Code of Academic Integrity, as explained in
the
Student Disciplinary Rules and Regulation. Violations of the rules will not be tolerated.
|