Course details


Machine Learning in Practice

WS 2021 Lucas Bechberger ONLINE
B.Sc modules:
CS-BWP-AI - Artificial Intelligence
CS-BWP-CL - (Computational) Linguistics
KOGW-WPM-CL - Computational Linguistics
KOGW-WPM-KI - Artificial Intelligence
M.Sc modules:
CC-MWP-AI - Artificial Intelligence
CC-MWP-CL - Computational Linguistics
CS-MWP-AI - Artificial Intelligence
CS-MWP-CL - (Computational) Linguistics

CS-BW - Bachelor elective course
CS-MW - Master elective course

In this seminar, we will look at the overall machine learning process from a software engineering perspective. We will cover both the main steps in a machine learning pipeline (from feature extraction over dimensionality reduction and hyperparameter optimization to incorporating the final model into an application) and relevant aspects from software engineering (including agile development methods, debugging, pair programming, and clean code). During the seminar, we will also make use of various development tools such as Trello, Slack, GitHub, and grid computing. This seminar will be held as an online block seminar. The main part of the seminar will take place from 04.10.2021 to 15.10.2021 and there will be a first introductory meeting in the week before (27.09. to 01.10. - to be determined via Stud.IP vote). The seminar will consist of joint sessions with seminar-style and lecture-style input and individual work in small groups (2 to 3 students), where the discussed concepts will be implemented as part of a simulated project. Grades will be determined based on quizzes about the theoretical contents and the code and documentation of the machine learning project simulated throughout the seminar. Prerequisites for this course include basic knowledge about machine learning (as for instance obtained from the "Machine Learning" lecture) and some programming experience in Python. Knowledge about software development in general (e.g., lecture "Informatik B") and about version control (git or svn) is certainly helpful, but not necessarily required. Also some background in Natural Language Processing may be helpful, since we will use text data as input to our machine learning system.