Machine learning is speeding up the discovery of new materials by combining the predictive power of quantum mechanics with fast simulations. Speaking at the AI for Science event, Professor Aron Walsh from Imperial College London and CUSP ai set out the core parts of materials artificial intelligence, including datasets, automation, models and optimisation. He highlighted constraints such as the availability of elements, toxicity and sustainable computing. Prof Walsh also described ways to represent chemical systems, including hand-built features, learned embeddings and foundation models.

Background: Automated Decision-Making in Policing

← Back to the news feed