Creative Applications of Machine Learning in Photography

Abstract

Artificial intelligence and deep learning hold considerable potential to be utilized for creative work, particularly in photography. This thesis focuses on the subject of machine learning via generative adversarial networks and its potential applications for creatives and artists in this field. In recent years machine learning has improved significantly as a result of big data and leaps in computer processing power. While the recent resurgence of novelty surrounding artificial intelligence has already raised substantial questions and criticisms in regards to its unforeseen consequences on our evolving relationship with technology in general, there is also a growing optimism aided by an increased democratization of access to machine learning. Visual artists from various backgrounds are learning to utilize artificial neural networks to conceptualize solutions for creative problems, as well as to augment and enhance their own work methods. This paper aims to examine the current state of artificial intelligence in photographic practices by exploring its technological innovations, creative tools, and applications. It will also address the problems of machine learning and AI-assisted art, and finally conclude with a prognosis for its future output potential.