Vertiefungsmodul Interdisziplinär VCA - Sender der Künste
Simon Grab
Nicholas Schärer
Assistenz Raffael Thielmann
4 CreditsBDE-VCA-V-I-3333.22H.001
Vertiefungsmodul Interdisziplinär VGD - Game Drawing II
Serge Pinkus
4 CreditsBDE-VGD-V-I-3333.22H.001
Vertiefungsmodul Interdisziplinär VIAD - Synthetic Normal
Paulina Zybinska
Yann Martins
4 CreditsBDE-VIAD-V-I-3333.06.22H.001
Vertiefungsmodul Interdisziplinär VID - Verpackungslösungen im Kontext von Nachhaltigkeit
Lisa Greuter, Véronique Rebetez
4 CreditsBDE-VID-V-I-3333.15.22H.001
Vertiefungsmodul interdisziplinär VSV - Infografik als animiertes GIF
Michael Stünzi
Daniel Röttele
Will Crook
4 CreditsBDE-VSV-V-I-3333.22H.001
Vertiefungsmodul Interdisziplinär VTI - Blut & Staub
Franziska Müller-Reissmann
4 CreditsBDE-VSD-V-I-3333-13.22H.001
Vertiefungsmodul Interdisziplinär VVK - Digital Publishing — Lost in Translation
Samuel Weidmann / Hauptleitung
Daniel Stutz / Hauptleitung
4 CreditsBDE-VVK-V-I-3333.22H.0011
Vertiefungsmodul Interdisziplinär VIAD - Synthetic Normal
Vertiefungsmodul Interdisziplinär VIAD - Synthetic Normal
Nummer und Typ | BDE-VIAD-V-I-3333.06.22H.001 / Moduldurchführung |
---|---|
Modul | Vertiefungsmodul Interdisziplinär VIAD - Synthetic Normal |
Veranstalter | Departement Design |
Leitung | Paulina Zybinska Yann Martins |
Zeit | Di 6. Dezember 2022 bis Fr 23. Dezember 2022 / 8:30–17 Uhr |
Anzahl Teilnehmende | 8 - 18 |
ECTS | 4 Credits |
Voraussetzungen | Willingness to work actively and self-motivated on projects Interest in new technologies At least some programming knowledge |
Lehrform | Class sessions include lectures, discussions, mentoring sessions, in-class presentations, assignments, independent study/investigation blocks and independent practical work on prototypes and final works. Projects are conducted in teams. |
Zielgruppen | Wahlpflichtmodul Bachelor Design, 3. Semester |
Lernziele / Kompetenzen | During the course we would like to explore a different approach to counter disobedient AI that revolves around the creation of the "Synthetic Normal". The goal is to use Machine Learning (ML) algorithms to create fake user profiles (bots) which match the criteria of “normalcy” i.e. based on the user data the system will have the ability to average the “out of the norm” qualities. During the course we will investigate the way social media platforms track users' posts, interactions with others, I/O devices, etc. by looking at debugging tools, web plugins and customised ML models. |
Inhalte | In the wake of the current landscape of corporate-based Artificial Intelligence (AI) systems, a growing number of artists, academics, designers, and computer scientists have started to question the use of this technology. Technically described by Joy Buolamwini and further discussed by Ruha Benjamin, AI has lately become synonymous with racial biases.To counter that Data Poisoning is used to inject crafted data and explore the ways by which AI systems could be hacked. The most notable example of data poisoning were "One pixel attacks" which were able to craft images where a single pixel could throw off most of image recognition AI softwares back in 2018. This and other techniques have been further applied to social media platforms as a way to obfuscate users' advertisable traits. While such forms of algorithmic defence might have been successful at first, companies behind the platforms keep on finding ways to mitigate such attacks. Furthermore, poisoned data has a drawback: data streams that are patternized become highly entropic, therefore asking for further analysis. |
Bibliographie / Literatur | Literature and links will be provided during the course and on the Paul Page (link to be updated). Leistungsnachweis Active and regular participation (min. 80% attendance) Project work In-class assignments presentations |
Leistungsnachweis / Testatanforderung | Aktive und regelmäßige Teilnahme (min. 80% Anwesenheit), Projektarbeit, In-class assignments, Präsentationen |
Termine | 06. - 23. Dezember 2022 (ohne Montage) |
Dauer | 3 Wochen |
Bewertungsform | Noten von A - F |
Bemerkung | The course is conducted in collaboration with Yann Patrick Martins, doctoral candidate of the Make/Sende Phd programme working at the Institute Experimental Design and Media Cultures (IXDM) as Researcher. Der Kurs wird in Englisch unterrichtet. Es kann Deutsch gesprochen werden, es sollte aber Englisch verstanden werden. |