Übung: 5944UE Machine Learning Lab - Details

Übung: 5944UE Machine Learning Lab - Details

Nachrichtenbox schließen
Sie sind nicht in Stud.IP angemeldet.

Allgemeine Informationen

Veranstaltungsname Übung: 5944UE Machine Learning Lab
Untertitel
Veranstaltungsnummer 5944UE
Semester SoSe 25
Aktuelle Anzahl der Teilnehmenden 29
erwartete Teilnehmendenanzahl 30
Heimat-Einrichtung Lehrstuhl für Data Science
beteiligte Einrichtungen Lehrstuhl für Informatik mit Schwerpunkt Verteilte Informationssysteme
Veranstaltungstyp Übung in der Kategorie Lehre (mit Prüfung)
Erster Termin Dienstag, 29.04.2025 13:00 - 17:00 Uhr, Ort: (IM) R 028
Art/Form
Teilnehmende
1.-3. Semester Master
Voraussetzungen
Visual Analytics or Network Science or Advanced Topics in Data Science
Leistungsnachweis
Portfolio exam consisting in the submission of the implementation code for selected machine learning algorithms plus documentation and the evaluation on a provided test-datasets. Students present their solution and results.
SWS
4
Literatur
Own Lecture Notes and selected publications.
Literature will be announced depending on the concrete topics.
Sonstiges
PO 2013:
Wahlpflichtmodul im Schwerpunkt Informations- und Kommunikationssysteme / compulsory elective module with a focus on information and communication systems
Wahlmodul im Schwerpunkt Algorithmik und Mathematische Modellierung / elective module with a focus on Algorithms and Mathematical Modelling
PO 2016:
Modulgruppe „Intelligente Technische Systeme“ / focus “intelligent technical systems”
ECTS-Punkte
6

Räume und Zeiten

(IM) R 028
Dienstag: 13:00 - 17:00, wöchentlich (12x)

Studienbereiche

Die Angaben zu den Anrechenbarkeiten an der FIM sind ohne Gewähr. Bitte beachten Sie die verbindliche Liste der Anrechenbarkeiten .

Modulzuordnungen

  • Universität Passau
    • Master Artificial Intelligence Engineering (Version WiSe 2021) (Hauptfach)
    • Master Computational Mathematics (Version SoSe 2018) (Hauptfach)
      • Abschluss MR CMT > Gesamtkonto MR CMT > Wahlpflichtbereich > Modulgruppe Data Analysis and Data Management and Programming
    • Master Informatik (Version SoSe 2016) (Hauptfach)
    • Master Mobile and Embedded Systems (Version WiSe 2016) (Hauptfach)

Kommentar/Beschreibung

Kenntnisse / Skills/Knowledge:
Students will acquire knowledge on implementation details of machine learning and optimization algorithms and how to realize them using numerical libraries in Python. Covered algorithms include supervised, unsupervised and semi-supervised algorithms like decision trees, support vector machines, Bayesian classifiers, hierarchical agglomerative clustering, Genetic algorithms etc. as well as optimization methods (e.g. stochastic gradient descent, AdaGrad)
Fähigkeiten / Abilities:
Students acquire the ability to implement machine learning algorithms from scratch using only numerical libraries. They will be able to evaluate their implementation and identify potential implementation errors.
Kompetenzen / Competencies:
Students acquire the skill to convert machine learning algorithms provided in a mathematical formulation or pseudo-code into concrete implementations. These skills include the implementation of performance metrics and the evaluation of the implemented algorithms without the help of third-party libraries.
During the semester, Students will be presented 6-10 different machine learning algorithms covering supervised, unsupervised, and supervised learning paradigms as well as different optimization methods. Examples are Decision Trees, Random Forests, Feedforward Neural Networks, Naive Bayes, Hierarchical Agglomerative Clustering, DB Scan, Support Vector Machine, Support Vector Regression, Stochastic Gradient Descent, AdaGrad etc.
During the lab sessions, students will have to implement those algorithms independently of each other using high-level programming languages, particularly Python, but without the help of any high-level library. Students will also have to develop corresponding evaluation metrics, like precision, recall, accuracy, average precision etc. and evaluate the algorithms based on standardized test data sets.

Anmeldemodus

Die Auswahl der Teilnehmenden wird nach der Eintragung manuell vorgenommen.

This is the waiting list.