Universität Wien FIND

301169 SE Applied Machine Learning for biological problems (2019W)

5.00 ECTS (3.00 SWS), SPL 30 - Biologie
Continuous assessment of course work

Registration/Deregistration

Details

max. 20 participants
Language: English

Lecturers

Classes (iCal) - next class is marked with N

Location: Seminar room 5.43 in the (former) University of Economics building, Augasse 2-6, Core A, 5th floor

Thursday 03.10. 15:00 - 16:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 10.10. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 17.10. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 31.10. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 07.11. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 14.11. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 21.11. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 28.11. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 05.12. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 12.12. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 09.01. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 16.01. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 23.01. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A
Thursday 30.01. 15:00 - 17:00 Seminarraum A5.43, UZA Augasse 2-6, 5.Stock Kern A

Information

Aims, contents and method of the course

Goals: Participants will learn basic concepts of machine learning, including working environments, several specific methods, and evaluation strategies. After the course, they should be able to decide, whether a given biological problem can be tackled with machine learning. They should be able to assess the quality of machine learning-based approaches in the scientific literature. Additionally, they should also be able to solve simple biological problems with machine learning,
Content:
Data science working environments (cloud infrastructures, Python, SciPy-stack, Jupyter notebooks, etc.)
Data preprocessing
Machine learning basics (data-driven methodology, evaluation strategies, underfitting/overfitting, train-test splits, cross-validation, etc.)
Methods:
Tree-based learning (e.g. decision trees)
Kernel methods (e.g. support vector machines)
Neural networks (e.g. deep learning)
Ensembles (e.g. gradient boosting)

Assessment and permitted materials

Weekly exercises (75 pts.)
Group project/competition (25 pts.)
Bonus points are possible through additional activities (TBA)

Minimum requirements and assessment criteria

Weekly exercises must be worked on individually. An additional larger project towards the end of the course can be worked on individually or in groups.
Attendance is encouraged, since concepts required for the exercises and the project are discussed in class, and we want the seminar to be interactive.
Participants must achieve at least 50 pts. by the end of the course.
Grading:
< 50.0: 5
< 62.5: 4
< 75.0: 3
< 87.5: 2
else: 1

Students need to bring their own laptop (we can provide a few, if required).
There are no strict prerequisites, but basic command line and Python skills are recommended.

Examination topics

Everything presented in the course.
Slides are and will be available online.

Reading list

Books:
Bishop: Pattern recognition and machine learning (2006)
Hastie: The elements of statistical learning (2009)
Goodfellow: Deep learning (2016) (free online version: https://www.deeplearningbook.org/)

Papers:
Relevant scientific literature will be introduced in the course.

Association in the course directory

MMEI II-1.2, MMEI II-2.2, MMB W-2, MMB W-3, PIK 2

Last modified: Fr 06.09.2019 14:36