Course data
Course name: Artificial Intelligence in Data Science
Neptun ID: BMETE15MF75
Responsible teacher: János Török
Department: Department of Theoretical Physics
Programme: Courses for Physicist MSc students
Course data sheet: BMETE15MF75
Requirements, Informations



  • The course will be both online and offline! Lectures will be recorded and posted to teams, practices will be partially recorded beforehand, but I will present it in person in the class, since some of the exercises require some explanation. During the practices the teams meeting will be running, and those online may ask questions using it. If I announce some hint or other general information I will turn on the microphone. The final project presentation can also be made using teams, so it is possible to complete the course fully online!
  • Please note that the course relies on the fact that everybody has good knowledge of python especially numpy.
  • I will use four different systems to communicate with you (sorry):
    • neptun: most official, final notes, important and prompt messages
    • moodle: submit everything here, evaluation will also take place here. For final grade use the formula below and not the one moodle supplies.
    • the official institute webpage, requirements are posted here (may change due to virus lock down). Lectures and homeworks are posted here.
    • Microsoft meet: I will try to record the lectures and post it here. All online consultation will take place here.



  • Moodle accounts will be created for each user, for which I will collect email addresses during the first class. If you cannot attend the first class you can send me an email, for this purpose. Moodle is available at Your username will be the same as your email name. Use the "Forgotten your username or password?" option to generate and email new password. Please use moodle system to submit homework.
  • You are supposed to make pairs, homeworks are individual work, but the practice work, data challenge and the final projects are supposed to be group works.

Subjects covered


Introduction to machine learning from a physicist's perspective, with the aim to understand how it works and less emphasis on tricks or parameter oprimization


  1. Image segmentation
  2. Decision tree
  3. Deep learning (from scratch in numpy)
  4. Higher level implementations (tensorflow, sklearn, keras)
  5. Convolutional neural networks
  6. Pre-trained models
  7. Data augmentation
  8. Textual data
  9. Sequential data
  10. Game models


  • 50% homework
  • 50% class work
  • Project presented and accepted


  • Homeworks (individual) 30%
  • Practice solutions (pair) 10%
  • Data challenge (pair) 20%
  • Project and its presentation (pair) 40%


  • In moodle


  1. In the moodle


  1. ea01.pdf
  2. ea02.pdf
  3. ea03.pdf
  4. ea04.pdf
  5. ea05.pdf
  6. ea06.pdf
  7. ea07.pdf
  8. ea08.pdf
  9. ea09.pdf
  10. ea10.pdf
  11. ea11.pdf

Old materials

Last modified: 08.08.2020