Symposium MT03-Accelerated Materials Discovery Through Data-Driven AI and Automation

The rapid development of automation, machine learning and artificial intelligence in the physical sciences has opened a new frontier for materials discovery. High-throughput calculations aided by machine learning can rapidly explore entire compositional spaces, while automated labs can discover and optimize synthetic routes with little human oversight. However, these approaches have highlighted fundamental challenges of the materials discovery process unique to autonomous systems. Key roadblocks include human-free interpretation of online characterisation to successfully identify novel materials, the difficulty of high-throughput computational screenings to reveal synthesizable materials in underexplored chemical spaces, and the simulation and synthetic processing of crystal imperfections and disorder. To speed up materials innovation and tackle these ongoing challenges, this symposium aims to bring together interdisciplinary researchers (materials, chemistry, engineering, computer science, statistics, and robotics) to discuss how data-driven approaches (AI algorithms and automated computations and synthesis) can accelerate the scientific workflow for materials discovery. The topics to be covered include generative AI for materials design, autonomous experimental and computational workflows, automated post-synthesis material identification, and how large-scale materials data can be leveraged. By bringing together researchers across a wide range of disciplines we aim to share recent advancements, facilitate knowledge exchange, and outline the opportunities and challenges in this rapidly evolving field.

artificial intelligence autonomous autonomous research Computing informatics machine learning predictive simulation