Both sides previous revision Previous revision Next revision | Previous revision |
what_kind_of_ai_do_we_want._the_case_of_generative_ai [2023/03/16 12:22] – [Thur. 16.03.2023] fstalder@zhdk.ch | what_kind_of_ai_do_we_want._the_case_of_generative_ai [2023/05/01 21:42] (current) – [Fri. 17.03.2023] fstalder@zhdk.ch |
---|
| |
=== Presentation From text to image with AI - when, how and why? === | === Presentation From text to image with AI - when, how and why? === |
| by [[https://dvstudies.net/2021/11/19/eva-cetinic/|Dr. Eva Cetinić]], Postdoctoral Fellow. [[https://dvstudies.net/|Digital Visual Studies]], UZH |
| |
Introduction to the concept of multimodality within deep learning - the “revolution of 2021” with multimodal foundation models (e.g. CLIP). Discussion of the various aspects and problems that arise from models being trained on hundreds million image-text pairs from the Internet (e.g. bias, cultural specificity, limitations of risk mitigation techniques, consent of content use and copyright issues, etc.). Discussion of the concept of “prompting”; the notion of similarity between word and image; the relation to art - using existing art, creating new “art”; the aesthetics of generated images - how it started and where it is going; the potential impact on the perception of images and media content. | Her research interests focus on studying new research methodologies rooted in the intersection of artificial intelligence and art history. Particularly, she is interested in exploring deep learning techniques for computational image understanding and multimodal reasoning in the context of visual art. |
| |
| **Abstract:** |
| |
| Introduction to the concept of multimodality within deep learning - the “revolution of 2021” with multimodal foundation models (e.g. CLIP). Discussion of the various aspects and problems that arise from models being trained on hundreds of million image-text pairs from the Internet (e.g. bias, cultural specificity, limitations of risk mitigation techniques, consent of content use and copyright issues, etc.). Discussion of the concept of “prompting”; the notion of similarity between word and image; the relation to art - using existing art, creating new “art”; the aesthetics of generated images - how it started and where it is going; the potential impact on the perception of images and media content. |
| |
* [[https://www.biennial.com/collaborations/the-next-biennial-should-be-curated-by-a-machine-experiment-aitnb|The Next Biennial Should be Curated by a Machine: Experiment AI-TNB]] | Liverpool Biennial of Contemporary Art (2021) | * [[https://www.biennial.com/collaborations/the-next-biennial-should-be-curated-by-a-machine-experiment-aitnb|The Next Biennial Should be Curated by a Machine: Experiment AI-TNB]] | Liverpool Biennial of Contemporary Art (2021) |
* https://ai.biennial.com | * https://ai.biennial.com |
| |
[[https://dvstudies.net/2021/11/19/eva-cetinic/|Dr. Eva Cetinić]], Postdoctoral Fellow. [[https://dvstudies.net/|Digital Visual Studies]], UZH | |
| |
Her research interests focus on studying new research methodologies rooted in the intersection of artificial intelligence and art history. Particularly, she is interested in exploring deep learning techniques for computational image understanding and multimodal reasoning in the context of visual art. | |
| |
* Eva Cetinic. [[https://arxiv.org/abs/2211.15271|The Myth of Culturally Agnostic AI Models]] (28 Nov 2022) | * Eva Cetinic. [[https://arxiv.org/abs/2211.15271|The Myth of Culturally Agnostic AI Models]] (28 Nov 2022) |
| |
* [[https://adam.harvey.studio/| Adam Harvey]] | * [[https://adam.harvey.studio/| Adam Harvey]] |
* [[https://refikanadol.com|Refik Anadol]] | * [[https://refikanadol.com|Refik Anadol]] |
* Christopher Kulendran Thomas ([[https://www.newgalerie.com/?page=artists&id=25|Being Human]], 2019 | [[https://www.berlinartlink.com/2022/12/16/christopher-kulendran-thomas-imagines-alternate-realities-in-another-world|Finesse]], 2022 | [[https://www.frontart.org/artists/kulendran-thomas|Dataset #1-4]])| [[https://vimeo.com/777019419?embedded=true&source=vimeo_logo&owner=19338498|Video, KW Berlin]] | * R.H. Lossin (2023). [[https://www.e-flux.com/criticism/527236/refik-anadol-s-unsupervised|Refik Anadol’s “Unsupervised”]]. e-flux (March 14). |
* Libbey Heaney | * Christopher Kulendran Thomas ([[https://www.newgalerie.com/?page=artists&id=25|Being Human]], 2019 | [[https://www.berlinartlink.com/2022/12/16/christopher-kulendran-thomas-imagines-alternate-realities-in-another-world|Finesse]], 2022 | [[https://www.frontart.org/artists/kulendran-thomas|Dataset #1-4]]| [[https://vimeo.com/777019419?embedded=true&source=vimeo_logo&owner=19338498|Video, KW Berlin]]) |
* Lauren Lee McCarthy | * Lauren Lee McCarthy ([[https://lauren-mccarthy.com/LAUREN|Lauren]], [[https://lauren-mccarthy.com/SOMEONE|Someone]], 2019) |
* Memo Atkin | * Memo Atkin. [[https://mixed-news.com/en/tiktoks-latest-beauty-ar-filter-is-indistinguishable-from-reality|TikTok’s latest beauty AR filter is indistinguishable from reality]]. 2023 |
* Moon Landing | * Kendrick Lamar - [[https://www.youtube.com/watch?v=uAPUkgeiFVY|The Heart Part 5]], 2022 |
| * Francesca Panetta, Halsey Burgund. [[https://moondisaster.org|Moon Disaster Speech]]. 2019 |
| |
=== Group Work II === | === Group Work II === |
| |
Place: ETHZ, LFW B3, Universitätsstrasse 2, 8092 Zürich | Place: ETHZ, LFW B3, Universitätsstrasse 2, 8092 Zürich |
| |
| [[https://news.uchicago.edu/story/uchicago-scientists-develop-new-tool-protect-artists-ai-mimicry|UChicago scientists develop new tool to protect artists from AI mimicry]]. Feb 14, 2023 | [[https://arxiv.org/pdf/2302.04222.pdf| arxiv.org PDF]] |
| |
=== Group work III === | === Group work III === |
| |
=== Group presentations === | === Group presentations === |
| |
| ====== additional resources ====== |
| Eryk Salvaggio. CRITICAL TOPICS: AI IMAGES. Syllabus for 26 classes. |
| |
| https://www.cyberneticforests.com/ai-images |