Contact | Basic information | Handouts | Previous semesters

• Syllabus
• Grading
• Regrading
• Academic integrity

Supervised machine learning (linear regression, logistic regression, simple neural networks, linear classification, SVM, kernel methods, decision trees, bagging and boosting). Machine learning theory (bias vs. variance, overfitting and underfitting, automated model selection, PAC learning, VC dimension). Unsupervised machine learning (clustering, PCA). Online learning and reinforcement learning.

Recommended literature:

• [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
  in library: I-INF-H-10
• [B] Christopher M. Bishop: Pattern Recognition and Machine Learning. Springer 2006
  in library: I-INF-B-38

The lectures will be loosely inspired by the recommended literature and will cover only some chapters. Lectures will be the primary source of information for the exam.

Grading

Homework assignments (3 x 10 points)	30%
Projects (one week before the exam)	30%
Final exam	40%
Bonus assignments during tutorials (5% each tutorial)

Final grade: A: 90+, B: 80+, C: 70+, D: 60+, E: 50+
To receive the final grade, you have to receive at least 50% points on the written exam.

Project instructions: here

All homeworks solutions must be your own work. It is not permitted to search internet and literature for homework solutions.

Regrading

All complaints about homework grades must be submitted in writing within two weeks of the time when the graded solutions are available, but no later than on the business day before the exam date. By examining your solution in more detail, the grade can be increased or decreased. Before submitting a claim, carefully read the relevant materials.

Academic Integrity

Copying homework assignments, projects and cheating on exams is a serious violation of academic integrity.

Cheating on homework assignments and projects include copying somebody else's work (i.e. classmate, internet, literature) and handing the work under your name, allowing somebody else to copy your work, or excessive collaboration. Cheating on exam includes use of unauthorized devices, as well as communication with others during the exam.

The standard penalty for cheating on homework assignment is a grade of -100%. Penalty for cheating on exam is Fx with no option of retaking the exam. Serious infraction will be reported to the faculty disciplinary committee.

With respect to homeworks, we encourage discussion between students and in groups. However, the solution you hand in must be your own and described in your own words. To avoid problems, do not keep any notes from such discussions and wait for several hours after the discussion before writing up your own solution.