Lehrende: Holger Prang
Veranstaltungsart: Seminar
Orga-Einheit: Fachübergreifende Studienangebote
Anzeige im Stundenplan: Q: Machine Learning
Anrechenbar für:
Semesterwochenstunden: 2
Standort: Hamburg
Unterrichtssprache: Englisch
Min. | Max. Anzahl Teilnehmer:innen: - | 25
Anmeldegruppe: Q-Studies Master
Leistungsnachweis:
Zusätzliche Informationen zu Terminen: Das Seminar findet an 3 Blockterminen Fr/Sa im CityScience Lab statt. Samstagtermine ggfs. außerhalb der HCU.
Beschreibung: Machine learning is increasingly super-powering todays digital services. As a subject it is taught mostly to computer scientists and engineers. Inspired by Gene Kogans "The Neural Aesthetic" this class will utilize machine learning techniques for the creative and artistic process. The course is meant as creative research with the goal to gain a simplified yet intuitive understanding of neural networks. The first event will provide an introduction to the theory and will be followed by two practical workshops for individual collection and experimentation with training data, training process and creation of images based on the training material with generative algorithms. The training material for the machine learning process will be selected by the participants, but the overall context will be the city and its possible futures. Examples material of generated satellite imagery, city and building plans, landscapes and facades will be presented and the code shared with the students.
Kontakt: holger.prang@hcu-hamburg.de
Module: Q-M-Mod-001, SF_M01, SF_M02
Sonstige Information: Prüfungsleistung: BSPO 2009: 5 ECTS Presentation & Exhibition at the end of the term BSPO 2015: 2,5 ECTS Presentation & Exhibition at the end of the term (working in groups of 2)