OCAD University Open Research Repository

Future Renaissance

Luginbuhl, Chris (2019) Future Renaissance. Masters thesis, OCAD University.

Item Type: Thesis
Creators: Luginbuhl, Chris
Abstract:

Name: Chris Luginbuhl
University Name: OCAD University
Thesis Title: Future Renaissance
Degree Title: MFA
Program Title: Digital Futures
Year of Defence: 2019

This document explores image making in collaboration with neural networks, particularly Generative Adversarial Networks (GANs). Research creation is used to develop a collaborative relationship with artificial intelligence (AI) which moves beyond the use of the computer as a simple tool in the image-making process. As a way of deepening this reflection, the images are developed using speculative fiction to imagine what intelligent machines’ creation myths might look like in the distant future, and this helps suggest how we might form AI in the present. GANs are found to help express visual ideas by providing a wealth of imagery and textural detail which can be modified with the selection of training data and transfer learning. The difficulty of training GANs can be mitigated by using other machine learning techniques such as object detection to gather training data, and by working with low-resolution imagery that reduces computational demands, increasing accessibility.

Date: 22 March 2019
Divisions: Graduate Studies > Digital Futures
Date Deposited: 13 May 2019 14:16
Last Modified: 13 May 2019 14:16
URI: http://openresearch.ocadu.ca/id/eprint/2524

Actions (login required)

Edit View Edit View