Neuro-Symbolic Generative Art: A Preliminary Study

TL;DR

This paper introduces a hybrid neuro-symbolic approach to generative art, combining neural networks with symbolic constraints, and demonstrates its increased perceived creativity through human evaluations.

Contribution

It proposes a novel neuro-symbolic generative art method and provides preliminary evidence of its enhanced creativity compared to purely symbolic approaches.

Findings

Subjects found neuro-symbolic artifacts more creative (61%)

Subjects perceived the process as more creative (82%)

Preliminary human studies support the approach's potential

Abstract

There are two classes of generative art approaches: neural, where a deep model is trained to generate samples from a data distribution, and symbolic or algorithmic, where an artist designs the primary parameters and an autonomous system generates samples within these constraints. In this work, we propose a new hybrid genre: neuro-symbolic generative art. As a preliminary study, we train a generative deep neural network on samples from the symbolic approach. We demonstrate through human studies that subjects find the final artifacts and the creation process using our neuro-symbolic approach to be more creative than the symbolic approach 61% and 82% of the time respectively.