Modeling Beauty ?



Machine learning may be a hot topic in the tech industry currently, but what exactly does it represent and what can it be used for? This talk will explain ML in an approachable, easy to grasp way applied to an interesting concept: Can we determine abstract concepts like beauty using machine learning? What is the appeal of being able to do so? Does it make a machine more human? Or does it expose the human experience as a list of 1s and 0s?

This talk explores these questions and concerns in regards to applying machine learning to classically artistic pursuits. Participants will first be exposed to an overview of human-defined beauty. Then, they will be able to examine a convolutional neural network model, pre-trained on labelled images, defined using Python and TensorFlow, and judge its ability to determine aesthetic merit for input images.