Dresden 2017 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

MM: Fachverband Metall- und Materialphysik

MM 55: Topical session: Data driven materials design - structure maps

MM 55.1: Talk

Thursday, March 23, 2017, 10:15–10:30, BAR 205

Materials discovery with artificial intelligence — •Gareth Conduit and Philipp Verpoort — Department of Physics, University of Cambridge, UK

We have developed a computational tool that employs deep learning with neural networks to discover new materials. The tool combines databases of experimental results with Density Functional Theory calculations to get high accuracy across a broad range of compositions. This enables us to propose materials that are most likely to fulfil multivariate targets. This holistic approach to materials design has allowed us to propose four new nickel-base alloys for use in jet engines, whose properties have been experimentally verified, new Lithium-ion battery cathode materials, and titanium alloys.

The neural network approach to materials modelling can also assess the integrity of materials data. We have exploited this capability to automatically validate and correct entries in a commercial metal alloy and polymer database.