Laboratories | Winter semester 2025/2026

Assessment rules

Throughout this course you will acquire the knowledge and skills to properly conduct scientific research in the field of machine learning, as well as to implement algorithms from scratch.

Programming tools:

• The tasks need to be implemented in Python programming language.
• The code can be written in regular scripts (.py) or in the notebook (.ipynb). Regular scripts are the recommended option.
• Tasks can be solved on university’s computers or on your own laptops/notebooks.
• The necessary programming libraries include:
  • numpy — handling data in the form of arrays, loading datasets, saving results
  • matplotlib — creating plots and visualizing results in figures
  • scikit-learn — tools for experimental evaluation (e.g. cross-validation), classifiers, other machine learning algorithms, data preprocessing (e.g. PCA)
  • scipy — additional tools (e.g. euclidean distance calculation), statistical tests
• You can use the code from previous laboratories for support. Using libraries’ documentation / stack overflow / blog posts is allowed. However, the use of LLMs (Copilot, GPT, Gemini etc.) is not allowed during the laboratory.

Grading:

1. Each laboratory contains 3 tasks from a given topic. Each task is worth exactly 1 point. It is necessary to complete task 1 before moving to task 2 and further.

2. Tasks will be made available at the beginning of the class.

3. Once you finish the tasks (or the class time is over) you need to present the results and be ready to answer any questions about the contents of the prepared scripts.

4. Before the laboratory is over, the results of your work must be submitted to ePortal course. Important:

   • The laboratories are scored according to the number of tasks that you have completed during the class. Make sure the tasks you have completed are checked. Once the class is over you can complete remaining tasks, but the number of points does not change.
   • The file submitted to ePortal should be a script (.py / .ipynb), not the figure.

5. First laboratories (L0) are introductory. The remaining 7 are graded (which results in the total of 21 points you can collect).

6. The points are collected for the entire semester. Depending on the sum of points, the final grade is given according to the table:

   Grade	5.0	4.5	4.0	3.5	3.0
   Percent	90%	80%	70%	60%	50%
   Points	18.9	16.8	14.7	12.6	10.5

7. It is not possible to correct the tasks after the laboratory.

8. Committing plagiarism results in failing the laboratory. Plagiarism is defined as submitting a code that the student did not write himself or code that the student cannot explain.

Attendance: