Visualizing Energy Landscapes through Manifold Learning
Benjamin W. B. Shires, Chris J. Pickard
Abstract
A machine-learning approach to visualizing high-dimensional energy landscapes provides a powerful tool for identifying stable arrangements of atoms that, in turn, can drive the discovery of new materials.
Topics & Concepts
Computer scienceNonlinear dimensionality reductionManifold (fluid mechanics)Energy (signal processing)Artificial intelligenceTheoretical computer sciencePhysicsDimensionality reductionQuantum mechanicsMechanical engineeringEngineeringNeural Networks and ApplicationsData Visualization and AnalyticsGenerative Adversarial Networks and Image Synthesis