Computational Methods for the Critical Simulation of Image-based Social Media Platforms

• In Event: ICA Interactive Paper/Poster Session I
  In Interactive Poster Session: Computational Methods Interactive Poster Session

Abstract

Social media platforms like Instagram are organised around creating, sharing and consuming images. Increasingly, platforms are engineering computer vision algorithms to classify, organise and act on their immense user-generated image databases. These algorithms are central to the advertising model and cultural impact of social media platforms, yet are not open to public scrutiny. Our interactive poster develops a computational framework that allows communication scholars ready access to a computational toolkit to simulate and interrogate the inner workings of these computer vision approaches. By building the image machines of social media platforms in the public domain, the project aims to enable critical understandings and analyses of how digital media shapes culture.

Authors

• Daniel Angus, University of Queensland

• Nicholas Carah, U of Queensland

• Adam Smith, The U of Queensland