Warning

This page is located in archive.
Go to the latest version of this course pages.
Go the latest version of this page.

** Winter semester 2019/2020 **

** Where and when:** computer lab KN:E-132 at Building E on Charles square, Thursday 12:45-14:15 and Tuesday 11:00-12:30

If you are new to CTU, see the checklist for visiting students.

**What can you expect:** The labs require you to implement learning and inference algorithms for a variety of classifiers. Your implementations will be tested with different pattern recognition tasks. Each week a new assignment is introduced at the beginning of the lab, and you are expected to complete the task during the submission period. The discussion at the beginning of the lab session will link the theory presented in the lectures to the practical task in the weekly assignments. The remaining time of the lab is devoted to individual interactions between students and teaching assistants. See the detailed rules below.

** What do we expect: ** Basic knowledge of Python (check the links in the first lab's text if you need a help with this).

** Important Links: **

** Teaching: **

There is a discussion forum administered for this course that can be used to solicit help for the assignments. It is monitored by the lab assistants and it is **the preferred form of communication** for giving assistance for the assignments since all students can see the question and answer threads. Please check the forum first if you have some confusion about an assignment.

Date | Topic | Test | |
---|---|---|---|

26.9. | introduction, work with python, simple example | ||

3.10. | bayesian decision task | ||

10.10. | non-bayesian tasks - the minimax task | ||

17.10. | MLE, MAP and Bayes parameter estimation | * | |

24.10. | non-parametrical estimates - parzen windows | ||

31.10. | logistic regression | ||

7.11. | exam questions | * | practice tasks |

14.11. | linear classifier - perceptron | ||

21.11. | support vector machines | ||

28.11. | support vector machines 2 | ||

5.12. | adaboost | * | |

12.12. | k-means clustering | ||

19.12. | convolutional neural networks | * | |

9.1. | zápočet, exam questions |

There will be a short test at the beginning of the labs denoted with *. The questions in the tests will refer to material presented in prior lectures.

- Delivering all finished assignments with at least half of them delivered within the two week assignment period (i.e. gaining at least 6 points).
- Attending of the tests at the beginning of the *-labs. The points from all tests except the worst are counted (when attending all four, otherwise all are counted), so a single bad result will not degrade your mark.

- A new assignment is given at each lab and the two week submission period for that assignment begins. The submission deadline is midnight of the day before the lab two weeks later.
- Students must work independently on the assignment during the week, but are encouraged to discuss implementation problems or misunderstandings with the lab assistants either during the labs, by using the forum, or via email or direct consultation during office hours.
- The submitted solution
**must**be original work. Please see the plagiarism policy. Students must not make solutions to assignments available to anyone else. Credit will not be assigned in the case of plagiarism. - The solutions are delivered through the Upload system.
- There are templates prepared for all assignments with function headers and boiler plate code. The templates are available in a git repository. Template improvements in form of a comment in forum or merge / pull requests on the gitlab are very appreciated.
- For successful submission the student is required to answer several questions related to the implementation of the assigned tasks. Unless all the questions are answered correctly, the points for the task are not granted.
**During the labs**- a test is given on the denoted weeks,
- a new assignment is introduced by a teaching assistant, and
- the remainder of time is used to for discussion, problem solving and
**bonus assignment evaluation**.

- A solution is awarded
**8 points**if it is delivered within one week of the assignment date, and**6 points**if submitted within two weeks. After that, a penalty of 2 points per additional day is applied (i.e.,**4, 2, 0 points**). - You are required to successfully submit
*all*assignments, even those which would be awarded 0 points because of missed deadlines to receive the credit (“zápočet”). - Some assignments contain a
**bonus task**. You can obtain up to**4 extra points**by solving this task. Show us the solutions of the bonus tasks during the labs. They will be evaluated individually. Bonus tasks have no deadlines. They can be delivered any time during the semester.

- The points awarded from the labs will be scaled to make 50% of the exam score (50 points). By working on the bonus tasks, one can gain even more than 50 points.
- The formula for the final score is following: (assignments+bonuses)/96*33 + tests/15*17, where 96 is to normalise for 12 assignments delivered within one week, and 15 for the maximum credits that can be obtained from the best three tests scores.
- With bonus points, a student may get more than 50 points, which will be considered if the student is on the borderline of a higher final grade.
- The extra points may also compensate to some degree lost points during the exam test. However, accumulating a high score in the labs does not guarantee a good final mark; you must perform adequately on the written exam. In the case of an exceedingly poor result on the final written exam, the final evaluation will be made accordingly.
- You can follow you actual score and compare your progress with others here (updated every Friday morning).

- If an absence is well justified, you will be given a chance to take a make-up test. The conditions of compensation will be dealt with individually.

courses/be5b33rpz/labs/start.txt · Last modified: 2019/12/18 16:27 by shekhole