Classification and Generation of Digital Marble Art: OR and Deep Learning

From IFORS Developing Countries Online Resources
Jump to: navigation, search

by: Gerhard Wilhelm Weber, Meltem Atay & Suryati Sitepu

OR and Art

- Generative adversarial networks (GANs) are types of neural networks which can be used to generate new data.

- they can be used to generate art pieces.

- We developed a new GAN framework specific to generate digital abstract art,

- we are introducing an open digital dataset for further research and collaboration opportunities.


- This idea is based on merging certain Support Vector Machines (SVMs) (invented by Drucker et al. [1997]) with convolution operation of Convolutional Neural Networks (CNNs) (proposed by LeCun et al. [1995]),

- kernels widely used for SVMs are linear, nonlinear and radial basis function (RBF) kernel methods tested for CNNs,

- we are proposing a general model which would be able to generate visually attractive art pieces without human supervision using OR methodology (Sack [2019]),

- our results are partially answering recent open questions about GAN framework proposed by (Odena [2019]).

link to material: