Christian Kerrigan: Chemotaxis of decanol system through glass fiber paper. Time series.
Christian Kerrigan introduces a series of projects from 200 Year Continuum, the overarching title to work comprising of collaborations in art, science, and technology. Each project utilizes time as a key component in creating its functionality and sustainability.
Website: www.200yearcontinuum.com
Martin Hanczyc: Developing Artificial Life through Science and the Arts
I will present an overview of my collaborative work with artists and designers over the past years. Basic research on soft matter physics has produced dynamical structures that challenge our perceptions of living systems. These systems have inspired diverse exhibitions, bringing ideas form Artificial Life to the public. The latest efforts in this direction have produced a new educational program that links students from arts, design, biology and engineering called ABRA (Artificial Biology, Robotics and Arts).
Jitka Cejkova: Droplets Kapky
A short presentation of the creation of art-movie from droplet experiments
Thomas Y. Chen: A Dual Neural Network Framework for Distinctly Stylistic Artistic Creation
Artificial neural networks (ANNs) are multilayered machine learning algorithms that aim to roughly simulate the human brain through a series of interconnected nodes. They are highly useful for interdisciplinary applications such as computer vision, which is the study of computers gaining high-level insights from imagery and video, including identifying, classifying, and semantically segmenting objects and people. However, a more recent and novel phenomenon in regards to ANNs is their ability to generate art. Generative adversarial networks (GANs), specifically, can create images (e.g. of human faces) that have never existed before but seem very realistic. At the same time, many of these images may not seem to be “art” like, in that they do not conform to a consistent authentic artistic style. Therefore, in this ongoing work, we propose a framework that consists of two neural networks: Net A, which is a generative model that generates a random landscape (e.g. grassland, beach, small town, etc.) based on training on images sourced from Google Images with the relevant queries. Subsequently, we have Net B which is the “styler” in that it takes the image outputted by Net B and styles it according to Claude Monet’s artistic style. While currently we only create a pipeline that applies Monet’s style to this pipeline, immediate future work includes incorporating other famous artistic styles such as Van Gogh, and demonstrating its efficacy among other images such as faces, rather than only landscapes. This work raises the question: how can we extrapolate this method to creating completely unique art? While in this case we are using the expertise of historical artists, how can we create a distinct style independent of these examples?