A Case of De-colonialist AttnGAN Text to ImageSynthesis
Date Issued
March 24, 2021
Author(s)
Abstract
At a time when AI-related technological infrastructure is becoming increasingly common in many aspects of life, this project attempts to demonstrate how bias in AI programming can affect an AI model’s interpretation of a dataset and how it can subsequently impact under-represented subjects. It pivots on an AttnGAN Text to Image synthesis pipeline as the means to pragmatically explore, articulate, and address affairs of bias in historical and digital colonialism. It zooms in on Cyprus, a geographic region that is underrepresented in AI programming, as seen through the lenses of 19th century colonialism, and intends to be the outset of a conceptualisation process about post-colonial Cyprus and how it may represent itself on own terms, and in relation to ongoing discourses on digital colonialism.
As such, it fosters a critical perspective on AI related research and, at the same time, it aspires to give agency to an underrepresented with AI milieux socio-cultural locus, resulting in a series of documents that document this process.
As such, it fosters a critical perspective on AI related research and, at the same time, it aspires to give agency to an underrepresented with AI milieux socio-cultural locus, resulting in a series of documents that document this process.
Subjects