SCIENCE AS ART
Exploration of chemical space by machine learning
Pascal Friederich, University of Toronto, Canada
The image depicts a genetic algorithm as it explores the space of small molecules in search of those that fulfill specific solubility and synthesizability criteria. Each point in space represents a different molecule, positioned according to its chemical structure. A neural network-based algorithm drives the genetic algorithm to explore unknown regions of the chemical space.
A Finalist from the Science as Art competition at the 2019 MRS Fall Meeting.
Copyright for all Science as Art images belongs to the Materials Research Society. To request permission to re-use the images, please contact Anita Miller.