2-INF-150: Machine learning (Strojové učenie) Winter 2018 Základné informácie |

Contacts | Basic information | Homework assigments | Lectures and handouts | Older versions of this course

Supervised learning (linear regression, logistic regression, simple neural networks, linear classification, SVM, kernel methods, decision tress, bagging and boosting). Theory of ML (overfitting, automated selection of models, PAC learning, VC dimension). Unsupervised learning (cluster, PCA). Reinforcement learning.

* Strojové učenie s učiteľom (lineárna regresia, logistická
regresia, jednoduché neurónové siete, lineárna klasifikácia, SVM,
kernelové metódy, rozhodovacie stromy, bagging a boosting). Teória
strojového učenia (preučenie, automatizovaný výber modelov, PAC
učenie, VC dimenzia). Strojové učenie bez učiteľa (zhlukovanie,
PCA). Učenie odmenou a trestom. *

**Recommended reading:**

- [GBC] Ian Goodfellow, Yoshua Bengio, Aaron Courville:
*Deep Learning.*MIT Press 2016 - [HTF] Trevor Hastie, Robert Tibshirani, Jerome Friedman:
*The Elements of Statistical Learning, 2nd ed.*Springer 2009

v knižnici: I-INF-H-10 - [B] Christopher M. Bishop:
*Pattern Recognition and Machine Learning.*Springer 2006

v knižnici: I-INF-B-38

Theoretical homeworks (2 of them, each for 7 points) | 14% |

Practical homeworks (8 of them, each for 4 points, best 4 are counted) | 16% |

Project (deadline 3 days before exam, at most until 31.1.2018) | 30% |

Exam | 40% |

Bonus points during semester (from tutorials) |

Grading scale: A: 90+, B: 80+, C: 70+, D: 60+, E: 50+

To get mark A-E it is needed to acquire at least 40% of points from exam.

Do homeworks and projects by yourself. It is forbidden to search for homeworks solution over the internet or in literature.

All appeals should be submitted **by two weeks** after homeworks is graded, but
at least one day before the exam day.
You grade might then change in both directions.

Cheating is bad, stinks and it will be punished, since it is serious violation of academic integrity

Cheating includes copying work from somebody else or from internet or literature and submitting it under your own name. It also includes allowing other person to copy your own work and also cooperation violating allowed limits. During exam it includes using forbided resources and communication with other people.

Cheating is typically punished by -100% grade for given part of work, downgrading by at least one level. During exam it means failure on the current exam. Serious violation would be reported to disciplinary commitee.

We allow discusion about homeworks between student during solving homeworks. But your solution should be written by yourself using your own words. To avoid problems, it is recommended to not take notes during group discussions and write down solution after a while (like three hours).

Maintained by 2-INF-150 personnel