Skip to content
Thesis Info
- LABS ID
- 00349
- Thesis Title
- Neural Networks In New Media Art
- Author
- Tim Barrass
- E-mail
- barrasstim AT gmail.com
- 2nd Author
- 3rd Author
- Degree
- Masters by research
- Year
- 2005
- Number of Pages
- 132
- University
- Melbourne University
- Thesis Supervisor
- Dr Peter Morse
- Supervisor e-mail
- peter AT petermorse.com.au
- Other Supervisor(s)
- Language(s) of Thesis
- English
- Department / Discipline
- Arts Faculty, School of Creative Arts
- Copyright Ownership
- Tim Barrass
- Languages Familiar to Author
- English
- URL where full thesis can be found
- Keywords
- art, media, neural, network, emergence, emergent, learning, agency, agent, self-organising, self-organisation, semiotic, analyse, analysis
- Abstract: 200-500 words
- This thesis presents a study of neural network based art. It discusses artificial neural networks as artistic media characterised by emergence, learning and agency. Based on these characteristics, a systemic-functional semiotic model for organising observations of neural net artworks is proposed. The model is applied to discuss example works including the author's own creative practice. The creative practice component of the study consists of software experiments with a population of neural net based agents that self-organise to generate digital “drawings” and animations. The process is described in the thesis along with the results in the form of colour prints and an accompanying DVD.