Researchers at artificial intelligence (AI) firm DeepMind used machine learning to simulate how atomic particles in a piece of glass respond to different temperatures and pressures.
The AI ran the software several times to account for all the various combinations of particles and neighbor particles, and to model how the entire piece of glass would react to different conditions.
The AI's predictions of initial particle movements under different pressures and temperatures achieved an average accuracy of 96%, which fell to 64% over longer time scales, but was still more accurate than current computer modeling techniques.
The researchers hope to use this AI to model traffic flow, treating cars as particles and using the same neighbor-particle concept to forecast vehicles' behavior in traffic jams.
From New Scientist
View Full Article - May Require Paid Subscription
Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA